LCAD - Contribuições do(a) usuário(a) [pt-br]

Equipe

2021-05-26T17:42:37Z

Rodrigo Berriel:

__TOC__

==Pesquisadores==

* [http://www.lcad.inf.ufes.br/team/index.php/Dr._Alberto_Ferreira_De_Souza Dr. Alberto Ferreira De Souza (Coordinator)]
* Dr. Claudine Badue
* Dr. Elias Oliveira
* Dr. Thiago Oliveira dos Santos
* Dr. Edilson de Aguiar (Egresso)

==Estudantes de Doutorado==

'''Professor Alberto'''
* Sabrina Siqueira Panceri
* Raphael Vivacqua Carneiro
* Vinicius Brito Cardoso
* Jairo Lucas de Moraes
* Rânik Guidolini
* Cayo Magno da Cruz Fontana

'''Professora Claudine'''
* Pedro Henrique Vieira de Oliveira Azevedo

'''Professor Elias'''
* Cristiano da Silveira Colombo
* Marcos Alécio Spalenza
* Wesley Pereira da Silva
*Jéssica Oliveira Brito (Coorientador de Doutorado)
*Jaimel de Oliveira Lima (Coorientador de Doutorado)
*Flávio Izo (Coorientador de Doutorado)

'''Professor Thiago'''
* Jacson Rodrigues Correia da Silva
* [https://rodrigoberriel.com Rodrigo Ferreira Berriel]
* Thiago Meireles Paixão

==Estudantes de Mestrado==

'''Professor Alberto'''
* Genilson de Morais Cruz
* Fabiano Rodrigues de Paula
* Marcos Thiago Piumbini
* Renan Mantuanelli de Aquino
* Sergio Costa Almeida
* Cézar Augusto Gobbo Passamani
* Marcelo Bringuenti Pedro

'''Professora Claudine'''
* Anderson Mozart Caetano dos Santos
* Caio Barata de Pinho
* David Pereira de Araújo
* Gabriel Andrade Nunes de Moraes
* Josias Alexandre Oliveira
* Thiago Gonçalves Cavalcante

'''Professor Elias'''
*James Alves da Silva Junior

'''Professor Thiago'''
*Letícia Carvalheiro Navarro
*Leonardo Paulucio
*Vinicius Ferraço Arruda
*Jean Pablo Vieira de Mello

==Estudantes de Graduação==

'''Professor Thiago'''
* Lucas Tabelini Torres

==Iniciação Científica==

'''Professor Elias'''
* Matheus Rizzi
* Gabriel Valdino

==Egressos==

'''Professor Alberto'''
* Kenyo Colnago dos Santos (Coorientador de Mestrado)
* João Paulo Angeli (Coorientador de Mestrado)
* Edilson Luiz do Nascimento (Coorientador de Mestrado)
* Fabio Daros de Freitas (Doutorado)
* Heráclito Amâncio Pereira Júnior (Doutorado)
* Filipe Wall Mutz (Doutorado)
* Lucas de Paula Veronese (Doutorado)
* Mariella Berger Andrade (Doutorado)
* Avelino Forechi Silva (Doutorado)
* Ricardo Emanuel Vaz Vargas (Graduação)
* Eduardo Max Amaro Amaral (Graduação)
* Hélcio Bezerra de Mello (Graduação)
* Luiz Eduardo Favalessa Peruch (Iniciação Científica)
* Daniel Cardoso (Iniciação Científica)
* Henrique Bertolo Selga (Iniciação Científica)
* Leonardo de Sena Cunha (Iniciação Científica)
* Lucas Catabriga Rocha (Iniciação Científica)
* Jorcy de Oliveira Neto (Iniciação Científica)
* Rafael Favaro Ferreira (Iniciação Científica)
* Renato Luiz de Freitas Cunha (Iniciação Científica)
* Helder Brito Nascimento (Iniciação Científica)
* Stiven Schwanz Dias (Iniciação Científica)
* Paulo Henrique Schmidt Castellani (Iniciação Científica)
* Thiago Thomes Coelho (Iniciação Científica)
* Melina Schneider Campo Dall'Orto (Iniciação Científica)
* Marcelo Bringuenti Pedro (Iniciação Científica)
* Thómas Jéfferson da Silva Teixeira (Mestrado)
* Rafael Correia Nascimento (Mestrado)
* Vinicius Brito Cardoso (Mestrado)
* Lauro José Lyrio Júnior (Mestrado)
* Tiago Alves DE Oliveira (Mestrado)
* Cayo Magno Da Cruz Fontana (Mestrado)
* Michael André Gonçalves (Mestrado)
* Vítor Barbirato Azevedo (Mestrado)
* André Gustavo Coelho de Almeida (Mestrado)
* Hélio Perroni Filho (Mestrado)
* Bruno Zanetti Melotti (Mestrado)
* Camilo Alves Carvalho (Mestrado)
* Felipe Thomaz Pedroni (Mestrado)
* Fernando Líbio Leite Almeida (Mestrado)
* Sotério Ferreira de Souza (Mestrado)
* Dijalma Fardim Júnior (Mestrado)
* Hallysson Oliveira (Mestrado)
* Sergio Nery Simões (Mestrado)
* Christian Daros de Freitas (Mestrado)
* Karin Satie Komati (Mestrado)
* Marta Magda Dornelles Bertoldi (Mestrado)
* Avelino Forechi Silva (Mestrado)

'''Professora Claudine'''

*Eduardo Max Amaro Amaral (Coorientador de Mestrado)
*Vinicius Cardoso Brito (Coorientador de Mestrado)
*Vítor Barbirato Azevedo (Coorientador de Mestrado)
*André Leal Tibério (Graduação)
*Jader Gomes Nascimento (Graduação)
*Josias Alexandre Oliveira (Graduação)
*Rickson Guidolini (Graduação)
*Vicente Bissoli Sessa (Graduação)
*Victor Nascimento Neves (Graduação)
*Állek Cezana Rajab (Iniciação Científica)
*Eduarda Costa Coppo (Iniciação Científica)
*Gabriel Hendrix (Iniciação Científica)
*Jader Gomes Nascimento (Iniciação Científica)
*Josias Alexandre Oliveira (Iniciação Científica)
*Lucas Grigoleto Scart (Iniciação Científica)
*Maurício Garcia Valli (Iniciação Científica)
*Miguel Rios Escóssia de Oliveira (Iniciação Científica)
*Rafael Correia Nascimento (Iniciação Científica)
*Rânik Guidolini (Iniciação Científica)
*Victor Nascimento Neves (Iniciação Científica)
*William de Oliveira Santos (Iniciação Científica)
*Caribe Zampirolli de Souza (Mestrado)
*Luan Ferreira Reis de Jesus (Mestrado)
*Ranik Guidolini (Mestrado)
*Rômulo Ramos Radaelli (Mestrado)
*Vicente Bissoli Sessa (Mestrado)

'''Professor Edilson'''
*Lucas de Paula Veronese (Coorientador de Doutorado)
*André Teixeira Lopes (Coorientador de Mestrado)
*Rodrigo Ferreira Berriel (Coorientador de Mestrado)
*Tiago Alves de Oliveira (Coorientador de Mestrado)
*Cayo Fontana (Coorientador de Mestrado)
*Lukas Taves (Graduação)
*Joao Paulo Moro Loureiro (Iniciação Científica)
*Raphael Assis (Iniciação Científica)
*Juliana Amorim Guimarães (Mestrado)
*Pedro Henrique Vieira de Oliveira Azevedo (Mestrado)

'''Professor Thiago'''

*Avelino Forechi Silva (Coorientador de Doutorado)
*Mariella Berger (Coorientador de Doutorado)
*Thómas Jéfferson da Silva Teixeira (Coorientador de Mestrado)
*Juliana Amorim Guimarães (Coorientador de Mestrado)
*Pedro Henrique Vieira de Oliveira Azevedo
*Filipe Wall Mutz (Coorientador de Mestrado)
*Lauro José Lyrio Júnior (Coorientador de Mestrado)
*Michael André Gonçalves (Coorientador de Mestrado)
*Renan Fricks dos Santos (Graduação)
*Gilmarllen Miotto (Graduação)
*David Morosini (Graduação)
*Ivo Nicchio (Graduação)
*Luiz Felipe Abreu Sousa (Graduação)
*Flávio Duarte (Graduação)
*Miguel Rios Escossia de Oliveira (Graduação)
*Rafael Correia Nascimento (Graduação)
*Gabriel Pietroluongo (Iniciação Científica)
*Franco Schmidt Rossi (Iniciação Científica)
*Vanderlei Vieira de Souza Filho (Iniciação Científica)
*João Paulo Araújo de Andrade (Iniciação Científica)
*Renan Sarcinelli (Mestrado)
*[https://rodrigoberriel.com Rodrigo Ferreira Berriel] (Mestrado)
*Lucas Luppi Amorim (Mestrado)
*Rafael Horimoto de Freitas (Mestrado)
*Marcus Vinicius Zavarez (Mestrado)
*Lucas Caetano Possatti (Mestrado)
*André Teixeira Lopes (Mestrado)
*Christian Baumberger (Mestrado)

Equipe

2021-05-26T17:36:19Z

Rodrigo Berriel: me adicionei como egresso de mestrado do Thiago

__TOC__

==Pesquisadores==

* [http://www.lcad.inf.ufes.br/team/index.php/Dr._Alberto_Ferreira_De_Souza Dr. Alberto Ferreira De Souza (Coordinator)]
* Dr. Claudine Badue
* Dr. Elias Oliveira
* Dr. Thiago Oliveira dos Santos
* Dr. Edilson de Aguiar (Egresso)

==Estudantes de Doutorado==

'''Professor Alberto'''
* Sabrina Siqueira Panceri
* Raphael Vivacqua Carneiro
* Vinicius Brito Cardoso
* Jairo Lucas de Moraes
* Rânik Guidolini
* Cayo Magno da Cruz Fontana

'''Professora Claudine'''
* Pedro Henrique Vieira de Oliveira Azevedo

'''Professor Elias'''
* Cristiano da Silveira Colombo
* Marcos Alécio Spalenza
* Wesley Pereira da Silva
*Jéssica Oliveira Brito (Coorientador de Doutorado)
*Jaimel de Oliveira Lima (Coorientador de Doutorado)
*Flávio Izo (Coorientador de Doutorado)

'''Professor Thiago'''
* Jacson Rodrigues Correia da Silva
* Rodrigo Ferreira Berriel
* Thiago Meireles Paixão

==Estudantes de Mestrado==

'''Professor Alberto'''
* Genilson de Morais Cruz
* Fabiano Rodrigues de Paula
* Marcos Thiago Piumbini
* Renan Mantuanelli de Aquino
* Sergio Costa Almeida
* Cézar Augusto Gobbo Passamani
* Marcelo Bringuenti Pedro

'''Professora Claudine'''
* Anderson Mozart Caetano dos Santos
* Caio Barata de Pinho
* David Pereira de Araújo
* Gabriel Andrade Nunes de Moraes
* Josias Alexandre Oliveira
* Thiago Gonçalves Cavalcante

'''Professor Elias'''
*James Alves da Silva Junior

'''Professor Thiago'''
*Letícia Carvalheiro Navarro
*Leonardo Paulucio
*Vinicius Ferraço Arruda
*Jean Pablo Vieira de Mello

==Estudantes de Graduação==

'''Professor Thiago'''
* Lucas Tabelini Torres

==Iniciação Científica==

'''Professor Elias'''
* Matheus Rizzi
* Gabriel Valdino

==Egressos==

'''Professor Alberto'''
* Kenyo Colnago dos Santos (Coorientador de Mestrado)
* João Paulo Angeli (Coorientador de Mestrado)
* Edilson Luiz do Nascimento (Coorientador de Mestrado)
* Fabio Daros de Freitas (Doutorado)
* Heráclito Amâncio Pereira Júnior (Doutorado)
* Filipe Wall Mutz (Doutorado)
* Lucas de Paula Veronese (Doutorado)
* Mariella Berger Andrade (Doutorado)
* Avelino Forechi Silva (Doutorado)
* Ricardo Emanuel Vaz Vargas (Graduação)
* Eduardo Max Amaro Amaral (Graduação)
* Hélcio Bezerra de Mello (Graduação)
* Luiz Eduardo Favalessa Peruch (Iniciação Científica)
* Daniel Cardoso (Iniciação Científica)
* Henrique Bertolo Selga (Iniciação Científica)
* Leonardo de Sena Cunha (Iniciação Científica)
* Lucas Catabriga Rocha (Iniciação Científica)
* Jorcy de Oliveira Neto (Iniciação Científica)
* Rafael Favaro Ferreira (Iniciação Científica)
* Renato Luiz de Freitas Cunha (Iniciação Científica)
* Helder Brito Nascimento (Iniciação Científica)
* Stiven Schwanz Dias (Iniciação Científica)
* Paulo Henrique Schmidt Castellani (Iniciação Científica)
* Thiago Thomes Coelho (Iniciação Científica)
* Melina Schneider Campo Dall'Orto (Iniciação Científica)
* Marcelo Bringuenti Pedro (Iniciação Científica)
* Thómas Jéfferson da Silva Teixeira (Mestrado)
* Rafael Correia Nascimento (Mestrado)
* Vinicius Brito Cardoso (Mestrado)
* Lauro José Lyrio Júnior (Mestrado)
* Tiago Alves DE Oliveira (Mestrado)
* Cayo Magno Da Cruz Fontana (Mestrado)
* Michael André Gonçalves (Mestrado)
* Vítor Barbirato Azevedo (Mestrado)
* André Gustavo Coelho de Almeida (Mestrado)
* Hélio Perroni Filho (Mestrado)
* Bruno Zanetti Melotti (Mestrado)
* Camilo Alves Carvalho (Mestrado)
* Felipe Thomaz Pedroni (Mestrado)
* Fernando Líbio Leite Almeida (Mestrado)
* Sotério Ferreira de Souza (Mestrado)
* Dijalma Fardim Júnior (Mestrado)
* Hallysson Oliveira (Mestrado)
* Sergio Nery Simões (Mestrado)
* Christian Daros de Freitas (Mestrado)
* Karin Satie Komati (Mestrado)
* Marta Magda Dornelles Bertoldi (Mestrado)
* Avelino Forechi Silva (Mestrado)

'''Professora Claudine'''

*Eduardo Max Amaro Amaral (Coorientador de Mestrado)
*Vinicius Cardoso Brito (Coorientador de Mestrado)
*Vítor Barbirato Azevedo (Coorientador de Mestrado)
*André Leal Tibério (Graduação)
*Jader Gomes Nascimento (Graduação)
*Josias Alexandre Oliveira (Graduação)
*Rickson Guidolini (Graduação)
*Vicente Bissoli Sessa (Graduação)
*Victor Nascimento Neves (Graduação)
*Állek Cezana Rajab (Iniciação Científica)
*Eduarda Costa Coppo (Iniciação Científica)
*Gabriel Hendrix (Iniciação Científica)
*Jader Gomes Nascimento (Iniciação Científica)
*Josias Alexandre Oliveira (Iniciação Científica)
*Lucas Grigoleto Scart (Iniciação Científica)
*Maurício Garcia Valli (Iniciação Científica)
*Miguel Rios Escóssia de Oliveira (Iniciação Científica)
*Rafael Correia Nascimento (Iniciação Científica)
*Rânik Guidolini (Iniciação Científica)
*Victor Nascimento Neves (Iniciação Científica)
*William de Oliveira Santos (Iniciação Científica)
*Caribe Zampirolli de Souza (Mestrado)
*Luan Ferreira Reis de Jesus (Mestrado)
*Ranik Guidolini (Mestrado)
*Rômulo Ramos Radaelli (Mestrado)
*Vicente Bissoli Sessa (Mestrado)

'''Professor Edilson'''
*Lucas de Paula Veronese (Coorientador de Doutorado)
*André Teixeira Lopes (Coorientador de Mestrado)
*Rodrigo Ferreira Berriel (Coorientador de Mestrado)
*Tiago Alves de Oliveira (Coorientador de Mestrado)
*Cayo Fontana (Coorientador de Mestrado)
*Lukas Taves (Graduação)
*Joao Paulo Moro Loureiro (Iniciação Científica)
*Raphael Assis (Iniciação Científica)
*Juliana Amorim Guimarães (Mestrado)
*Pedro Henrique Vieira de Oliveira Azevedo (Mestrado)

'''Professor Thiago'''

*Avelino Forechi Silva (Coorientador de Doutorado)
*Mariella Berger (Coorientador de Doutorado)
*Thómas Jéfferson da Silva Teixeira (Coorientador de Mestrado)
*Juliana Amorim Guimarães (Coorientador de Mestrado)
*Pedro Henrique Vieira de Oliveira Azevedo
*Filipe Wall Mutz (Coorientador de Mestrado)
*Lauro José Lyrio Júnior (Coorientador de Mestrado)
*Michael André Gonçalves (Coorientador de Mestrado)
*Renan Fricks dos Santos (Graduação)
*Gilmarllen Miotto (Graduação)
*David Morosini (Graduação)
*Ivo Nicchio (Graduação)
*Luiz Felipe Abreu Sousa (Graduação)
*Flávio Duarte (Graduação)
*Miguel Rios Escossia de Oliveira (Graduação)
*Rafael Correia Nascimento (Graduação)
*Gabriel Pietroluongo (Iniciação Científica)
*Franco Schmidt Rossi (Iniciação Científica)
*Vanderlei Vieira de Souza Filho (Iniciação Científica)
*João Paulo Araújo de Andrade (Iniciação Científica)
*Renan Sarcinelli (Mestrado)
*Rodrigo Ferreira Berriel (Mestrado)
*Lucas Luppi Amorim (Mestrado)
*Rafael Horimoto de Freitas (Mestrado)
*Marcus Vinicius Zavarez (Mestrado)
*Lucas Caetano Possatti (Mestrado)
*André Teixeira Lopes (Mestrado)
*Christian Baumberger (Mestrado)

Wiki do LCAD

2018-04-17T18:21:42Z

Rodrigo Berriel: add pagina

Bem-vindo à página de documentação do LCAD.

O objetivo deste Wiki é documentar todas as ferramentas, procedimentos de instalação e de suporte oferecidos no Laboratório de Computação de Alto Desempenho (LCAD).

Abaixo você pode conferir alguns roteiros:

:*[[Lista de Aquisições]]
:*[[Lista de Ocupação de Endereços IP da Rede LCAD]]
:*[[Carmen Robot Navigation Toolkit]]
:*[[Máquina Associadora de Eventos - MAE]]
:*[[Protocolo de Concessão de Contas no LCAD]]
:*[[Processo de Criação de Contas no LCAD]]
:*[[Hardwares e Serviços]]
:*[[:Media:Tutorial wiki.pdf|Guia de Utilização do Wiki LCAD]]
:*[[Guia de Utilização do Doxygen]]
:*[[Micro Manual de Utilização do Cluster]]
:*[[Processamento Paralelo]]
:*[[Como Funciona o Backup do LCAD]]
:*[[Como Utilizar a Rede Wireless no LCAD]]
:*[[Uso Básico do Subversion]]
:*[[Instalações Úteis]]
:*[[Dicas e Configurações]]
:*[[Construção de pacotes RPM|Como construir pacotes RPM - Exemplo básico]]
:*[[Telefones Úteis]]
:*[[Como preparar café]]
:*[[Planejamento de Atividades]]
:*[[Apresentações]]
:*[[Identidade Visual]]
:*[[Material de Consumo Específico]]
:*[[Como criar branches no SVN]]
:*[[Roteiro de Instalação do Fedora nos Blades para o Carro]]
:*[[Roteiro de como configurar NIS no FEDORA 17]]
:*[[Roteiro de Instalação do Apache e SVN com autenticação NIS]]
:*[[Instalar o Confirm Account no Wiki]]
:*[[Configurar SSH sem a necessidade de usar senha]]
:*[[Criando um repositório no SVN]]
:*[[Instalando/Conectando via VNC no Carro]]
:*[[Criando servidor NTP e atualizando as máquinas pela rede]]
:*[[Adicionando latência com tc no Linux]]
:*[[Criando Kernel RT linux]]
:*[[Criando Kernel RT Ubuntu 14.04]]
:*[[Clonar e Restaurar HD com Clonezilla]]
:*[[Como usar os pacotes ROS visual search thin e visual search webcam]]
:*[[Como usar o pacote ROS cyton]]
:*[[Criando mapas Carmen]]
:*[[Repetidor Wireless]]
:*[[Servidor DHCP]]
:*[[Servidor DHCP Wireless]]
:*[[Procedimento para Limpar Mapas do Carmen]]
:*[[Como obter GUID das câmeras PointGrey]]
:*[[Estrutura de Rede do LCAD]]
:*[[Roteiro de Instalação Python/Theano]]
:*[[Como Cadastrar o MAC na rede do LCAD]]
:*[[Como adicionar um artigo]]

Para uma listagem mais completa dos artigos contidos neste Wiki, você pode [[Special:Categories|navegar pelas categorias.]]

Como adicionar um artigo

2018-04-17T18:14:29Z

Rodrigo Berriel: minor

Siga o passo-a-passo abaixo para adicionar uma página do seu artigo na lista de publicações do LCAD na Wiki. Para seguir o passo-a-passo abaixo, você precisará estar logado.

== Passo-a-passo ==

1. Faça uma busca pelo título do seu artigo na caixa de busca da Wiki do LCAD. Nesse tutorial, usaremos o título "My Awesome Paper" como exemplo;

2. Como resultado, você verá algo como: "'''Criar a página "[http://www.lcad.inf.ufes.br/wiki/index.php?title=My_Awesome_Paper&action=edit&redlink=1 My Awesome Paper]" nesta wiki!'''" ([http://www.lcad.inf.ufes.br/wiki/index.php?search=My+Awesome+Paper&title=Especial%3ABusca&go=Ir veja aqui]);

3. Clicando no título do artigo (e.g., clique no título no item anterior), você será redirecionado para um página onde poderá colocar o conteúdo da página do seu artigo. Veja abaixo sugestões de conteúdo para colocar em uma página de artigo na Wiki;

4. Ao terminar a edição, clique em "Salvar página". Pronto: sua página está criada. Se você tiver adicionado a categoria "Publicações" (conforme sugerido), você já poderá ver seu artigo sendo listado nessa página: [http://www.lcad.inf.ufes.br/wiki/index.php/Categoria:Publicações Publicações];

5. Além disso, é importante que você adicione o seu artigo na página de [http://www.lcad.inf.ufes.br/wiki/index.php/Publicações_do_LCAD Publicações do LCAD]. Para isso, acesse esse [http://www.lcad.inf.ufes.br/wiki/index.php/Publicações_do_LCAD link] e clique em "[http://www.lcad.inf.ufes.br/wiki/index.php?title=Publica%C3%A7%C3%B5es_do_LCAD&action=edit Editar]" no canto superior direito.

6. Ao editar, siga o padrão dos outros artigos na lista e adicione uma linha com o seu no final. O padrão atual é:

[[My Awesome Paper]], 20XX, Nome do Periódico ou Conferência

7. Após adicionar o seu artigo, clique em "Salvar página" e pronto! Acabou. Seu artigo já está disponível na Wiki do LCAD. Agora é só divulgar!

'''Dica:''' ''antes de clicar em "Salvar página", sempre clique em "Mostrar previsão" para saber como está ficando. Faça isso para saber como está ficando a formatação da página, por exemplo.''

== O que colocar na página do artigo? ==

A página do artigo é sua e você tem total liberdade para colocar as informações que desejar. Abaixo, são apenas algumas recomendações:

* Comece adicionando a página na categoria de Publicações:

<nowiki>[[category:Publicações]]</nowiki>

* Após isso, adicione os autores e ponteiros para a versão oficial e para uma versão de acesso gratuito:

<nowiki>Authors: Name1, Name2, Name3

DOI: link to DOI (e.g., [http://dx.doi.org/10.XXXX/YYYY 10.XXXX/YYYY])

PDF: link to ArXiV, Researchgate, etc.</nowiki>

* Adicione o resumo (abstract) do artigo:

<nowiki>== Abstract ==

[[Arquivo:my-awesome-graphical-abstract.png|x350px]]

My Awesome Abstract</nowiki>

* Adicione o BibTeX;

<nowiki>== BibTeX ==

@article{name20XXjournal,
author = {...},
journal = {...},
...
}</nowiki>

* Adicione outros itens de interesse (e.g., link para materiais, GitHub, vídeos).

'''Dica:''' ''você pode acessar páginas de outros artigos (e.g., [[Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach|1]], [[Ego-Lane Analysis System|2]], [[Deep Learning Based Large-Scale Automatic Satellite Crosswalk Classification|3]]), clicar em "Editar" e copiar a formatação para usar como base para o seu.''

Como adicionar um artigo

2018-04-17T17:53:11Z

Rodrigo Berriel: versão inicial do tutorial

Publicações do LCAD

2018-04-14T18:38:00Z

Rodrigo Berriel: Fix ego-lane link

[[category:Publicações]]
== [http://www.lcad.inf.ufes.br/~alberto/pesquisadores-lcad-producao Publicações do LCAD desde 1996] ==

== Papers Summaries ==

[[A Particle Filter-based Lane Marker Tracking Approach using a Cubic Spline Model]], 2015, SIBGRAPI

[[A Facial Expression Recognition System Using Convolutional Networks]], 2015, SIBGRAPI

[[Facial Expression Recognition with Convolutional Neural Networks: Coping with Few Data and the Training Sample Order]], 2016, Pattern Recognition International Journal

[[Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach]], 2017, Computers & Graphics

[[Deep Learning Based Large-Scale Automatic Satellite Crosswalk Classification]], 2017, IEEE Geoscience and Remote Sensing Letters

[[Ego-Lane Analysis System|Ego-Lane Analysis System (ELAS): Dataset and Algorithms]], 2017, Image and Vision Computing

Publicações do LCAD

2018-04-14T18:36:07Z

Rodrigo Berriel: Add Ego-Lane paper na lista

[[category:Publicações]]
== [http://www.lcad.inf.ufes.br/~alberto/pesquisadores-lcad-producao Publicações do LCAD desde 1996] ==

== Papers Summaries ==

[[A Particle Filter-based Lane Marker Tracking Approach using a Cubic Spline Model]], 2015, SIBGRAPI

[[A Facial Expression Recognition System Using Convolutional Networks]], 2015, SIBGRAPI

[[Facial Expression Recognition with Convolutional Neural Networks: Coping with Few Data and the Training Sample Order]], 2016, Pattern Recognition International Journal

[[Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach]], 2017, Computers & Graphics

[[Deep Learning Based Large-Scale Automatic Satellite Crosswalk Classification]], 2017, IEEE Geoscience and Remote Sensing Letters

[[Ego-Lane Analysis System (ELAS): Dataset and Algorithms]], 2017, Image and Vision Computing

Ego-Lane Analysis System

2018-03-01T12:10:10Z

Rodrigo Berriel: bug double-curly-braces

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo Berriel], [http://people.mpi-inf.mpg.de/~edeaguia/ Edilson de Aguiar], [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

'''Image and Vision Computing''': [https://doi.org/10.1016/j.imavis.2017.07.005 10.1016/j.imavis.2017.07.005]

== Abstract ==

[[Arquivo:Elas-graphical-abstract.png|x350px]]

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines, Kalman filter and Particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e. lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.

== Demonstration Video ==

Demonstration Video of ELAS, as proposed by the time of submission

[[Arquivo:Elas-video1.png|x200px|link=http://bit.ly/SI-AUTO-VISION_DEMO-VIDEO]]

ELAS was weakly integrated into IARA (our autonomous vehicle). The video below shows ELAS performing on IARA (without tuning any parameter).

[[Arquivo:Elas-carmen.png|x200px|link=https://www.youtube.com/watch?v=R5wdPJ4ZI5M]]

== Source-Code ==
Available [https://github.com/rodrigoberriel/ego-lane-analysis-system here]

== Dataset ==
To request access to the datasets, read the instructions [https://github.com/rodrigoberriel/ego-lane-analysis-system/blob/master/datasets/ here].

It contains 22 scenes with a total of 17,092 frames. In 98.54% of the frames, at least one of the sides has white lane markings, while 12.88% of them are yellow. In 5.18% of the frames at least one side has no lane markings. This numbers include frames containing transitions of lane markings type, i.e. images with two lane markings type simultaneously on the same side. Lane markings type transitions are present in 1,854 frames (10.85%). Lane change maneuver are being performed in 5.42% of this dataset, where 64.72% is from right to left and 35.28% in the opposite direction. There are intersections in 2.11% of the frames. In 7.28% of all images, there is at least one pavement marking. 33.92% of the pavement markings are crosswalks and 15.27% are annotated as unknown, i.e. they are of none of the classes of interest and 50.81% comprises arrows and stop lines. There is at least one adjacent lane in 50.34% and 72.14% of the frames for each side, right and left respectively.

[https://drive.google.com/open?id=1sIgybW-V3mrJ_CkjiC_IS2zEZjDJ5Pm2 Download a short description of each scene (with images).]

Some samples below:

[[Arquivo:Elas-dataset-samples.png|x300px]]

== BibTeX ==

@article{berriel2017imavis,
Author = {Rodrigo F. Berriel and Edilson de Aguiar and Alberto F. de Souza and Thiago Oliveira-Santos},
Title = {{Ego-Lane Analysis System (ELAS): Dataset and Algorithms}},
Journal = {Image and Vision Computing},
Volume = {68},
Pages = {64--75}
Year = {2017},
DOI = {10.1016/J.IMAVIS.2017.07.005},
ISSN = {0262-8856},
}

Ego-Lane Analysis System

2018-03-01T12:06:58Z

Rodrigo Berriel: correção link e bibtex

Publicações do LCAD

2017-10-10T16:46:19Z

Rodrigo Berriel:

Deep Learning Based Large-Scale Automatic Satellite Crosswalk Classification

2017-10-10T16:45:47Z

Rodrigo Berriel: Criou página com 'category:Publicações Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], André Teixeira Lopes, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferre...'

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], André Teixeira Lopes, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

DOI: [http://dx.doi.org/10.1109/LGRS.2017.2719863 10.1109/LGRS.2017.2719863]

PDF: [https://www.researchgate.net/publication/318010149_Deep_Learning_Based_Large-Scale_Automatic_Satellite_Crosswalk_Classification]

''Published in [http://ieeexplore.ieee.org/document/7979607 IEEE Geoscience and Remote Sensing Letters]''

== Abstract ==

[[Arquivo:Overview-grsl.png|x350px]]

High-resolution satellite imagery have been increasingly used on remote sensing classification problems. One of the main factors is the availability of this kind of data. Even though, very little effort has been placed on the zebra crossing classification problem. In this letter, crowdsourcing systems are exploited in order to enable the automatic acquisition and annotation of a large-scale satellite imagery database for crosswalks related tasks. Then, this dataset is used to train deep-learning-based models in order to accurately classify satellite images that contains or not zebra crossings. A novel dataset with more than 240,000 images from 3 continents, 9 countries and more than 20 cities were used in the experiments. Experimental results showed that freely available crowdsourcing data can be used to accurately (96.78%) train robust models to perform crosswalk classification on a global scale.

== Source-code and Models ==

Source-code and Models available on [https://github.com/rodrigoberriel/satellite-crosswalk-classification GitHub].

== BibTeX ==

@ARTICLE{berriel2017grsl,
author = {Rodrigo Ferreira Berriel and André Teixeira Lopes and Alberto Ferreira de Souza and Thiago Oliveira-Santos},
title = {Deep Learning Based Large-Scale Automatic Satellite Crosswalk Classification},
journal = {IEEE Geoscience and Remote Sensing Letters},
issn = {1545-598X}
volume = {14}
number = {9}
year = {2017},
month = {Sept},
pages = {1513-1517},
doi = {10.1109/LGRS.2017.2719863}
}

Arquivo:Overview-grsl.png

2017-10-10T16:38:20Z

Rodrigo Berriel:

Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach

2017-10-10T16:27:59Z

Rodrigo Berriel:

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], Franco Schmidt Rossi, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

DOI: [http://dx.doi.org/10.1016/J.CAG.2017.08.004 10.1016/J.CAG.2017.08.004]

PDF: [https://www.researchgate.net/publication/319194310_Automatic_Large-Scale_Data_Acquisition_via_Crowdsourcing_for_Crosswalk_Classification_A_Deep_Learning_Approach]

''Published in [https://www.journals.elsevier.com/computers-and-graphics/ Computers & Graphics]''

== Abstract ==

[[Arquivo:Crosswalk-streetview-graphical-abstract.png|x350px]]

Correctly identifying crosswalks is an essential task for the driving activity and mobility autonomy. Many crosswalk classification, detection and localization systems have been proposed in the literature over the years. These systems use different perspectives to tackle the crosswalk classification problem: satellite imagery, cockpit view (from the top of a car or behind the windshield), and pedestrian perspective. Most of the works in the literature are designed and evaluated using small and local datasets, i.e. datasets that present low diversity. Scaling to large datasets imposes a challenge for the annotation procedure. Moreover, there is still need for cross-database experiments in the literature because it is usually hard to collect the data in the same place and conditions of the final application. In this paper, we present a crosswalk classification system based on deep learning. For that, crowdsourcing platforms, such as OpenStreetMap and Google Street View, are exploited to enable automatic training via automatic acquisition and annotation of a large-scale database. Additionally, this work proposes a comparison study of models trained using fully-automatic data acquisition and annotation against models that were partially annotated. Cross-database experiments were also included in the experimentation to show that the proposed methods enable use with real world applications. Our results show that the model trained on the fully-automatic database achieved high overall accuracy (94.12%), and that a statistically significant improvement (to 96.30%) can be achieved by manually annotating a specific part of the database. Finally, the results of the cross-database experiments show that both models are robust to the many variations of image and scenarios, presenting a consistent behavior.

== Videos ==

See the IARA, GOPRO, and NIGHT dataset videos [https://www.youtube.com/playlist?list=PLm8amuguiXiIbZNQHj0BH1AJddnnql356 here].

== Source-code and Models ==

Available soon: [https://github.com/rodrigoberriel/streetview-crosswalk-classification GitHub].

== BibTeX ==

@ARTICLE{Berriel2017cag,
author = {Rodrigo Ferreira Berriel and Franco Schmidt Rossi and Alberto Ferreira de Souza and Thiago Oliveira-Santos},
journal = {Computers & Graphics},
issn = {0097-8493}
title = {Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach},
volume = {68}
year = {2017},
month = {Nov},
pages = {32-42},
doi = {10.1016/J.CAG.2017.08.004}
}

Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach

2017-10-10T16:25:26Z

Rodrigo Berriel: /* BibTeX */

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], Franco Schmidt Rossi, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

DOI: [http://dx.doi.org/10.1016/J.CAG.2017.08.004 10.1016/J.CAG.2017.08.004]

PDF: [https://www.researchgate.net/publication/319194310_Automatic_Large-Scale_Data_Acquisition_via_Crowdsourcing_for_Crosswalk_Classification_A_Deep_Learning_Approach]

''Published in [https://www.journals.elsevier.com/computers-and-graphics/ Computers & Graphics]''

== Abstract ==

[[Arquivo:Crosswalk-streetview-graphical-abstract.png|x350px]]

Correctly identifying crosswalks is an essential task for the driving activity and mobility autonomy. Many crosswalk classification, detection and localization systems have been proposed in the literature over the years. These systems use different perspectives to tackle the crosswalk classification problem: satellite imagery, cockpit view (from the top of a car or behind the windshield), and pedestrian perspective. Most of the works in the literature are designed and evaluated using small and local datasets, i.e. datasets that present low diversity. Scaling to large datasets imposes a challenge for the annotation procedure. Moreover, there is still need for cross-database experiments in the literature because it is usually hard to collect the data in the same place and conditions of the final application. In this paper, we present a crosswalk classification system based on deep learning. For that, crowdsourcing platforms, such as OpenStreetMap and Google Street View, are exploited to enable automatic training via automatic acquisition and annotation of a large-scale database. Additionally, this work proposes a comparison study of models trained using fully-automatic data acquisition and annotation against models that were partially annotated. Cross-database experiments were also included in the experimentation to show that the proposed methods enable use with real world applications. Our results show that the model trained on the fully-automatic database achieved high overall accuracy (94.12%), and that a statistically significant improvement (to 96.30%) can be achieved by manually annotating a specific part of the database. Finally, the results of the cross-database experiments show that both models are robust to the many variations of image and scenarios, presenting a consistent behavior.

== Videos ==

See the IARA, GOPRO, and NIGHT dataset videos [https://www.youtube.com/playlist?list=PLm8amuguiXiIbZNQHj0BH1AJddnnql356 here].

== Source-code and Models ==

Source-code and models available: [https://github.com/rodrigoberriel/satellite-crosswalk-classification GitHub].

== BibTeX ==

@ARTICLE{Berriel2017cag,
author = {Rodrigo Ferreira Berriel and Franco Schmidt Rossi and Alberto Ferreira de Souza and Thiago Oliveira-Santos},
journal = {Computers & Graphics},
issn = {0097-8493}
title = {Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach},
volume = {68}
year = {2017},
month = {Nov},
pages = {32-42},
doi = {10.1016/J.CAG.2017.08.004}
}

Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach

2017-10-10T16:25:08Z

Rodrigo Berriel: /* BibTeX */

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], Franco Schmidt Rossi, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

DOI: [http://dx.doi.org/10.1016/J.CAG.2017.08.004 10.1016/J.CAG.2017.08.004]

PDF: [https://www.researchgate.net/publication/319194310_Automatic_Large-Scale_Data_Acquisition_via_Crowdsourcing_for_Crosswalk_Classification_A_Deep_Learning_Approach]

''Published in [https://www.journals.elsevier.com/computers-and-graphics/ Computers & Graphics]''

== Abstract ==

[[Arquivo:Crosswalk-streetview-graphical-abstract.png|x350px]]

Correctly identifying crosswalks is an essential task for the driving activity and mobility autonomy. Many crosswalk classification, detection and localization systems have been proposed in the literature over the years. These systems use different perspectives to tackle the crosswalk classification problem: satellite imagery, cockpit view (from the top of a car or behind the windshield), and pedestrian perspective. Most of the works in the literature are designed and evaluated using small and local datasets, i.e. datasets that present low diversity. Scaling to large datasets imposes a challenge for the annotation procedure. Moreover, there is still need for cross-database experiments in the literature because it is usually hard to collect the data in the same place and conditions of the final application. In this paper, we present a crosswalk classification system based on deep learning. For that, crowdsourcing platforms, such as OpenStreetMap and Google Street View, are exploited to enable automatic training via automatic acquisition and annotation of a large-scale database. Additionally, this work proposes a comparison study of models trained using fully-automatic data acquisition and annotation against models that were partially annotated. Cross-database experiments were also included in the experimentation to show that the proposed methods enable use with real world applications. Our results show that the model trained on the fully-automatic database achieved high overall accuracy (94.12%), and that a statistically significant improvement (to 96.30%) can be achieved by manually annotating a specific part of the database. Finally, the results of the cross-database experiments show that both models are robust to the many variations of image and scenarios, presenting a consistent behavior.

== Videos ==

See the IARA, GOPRO, and NIGHT dataset videos [https://www.youtube.com/playlist?list=PLm8amuguiXiIbZNQHj0BH1AJddnnql356 here].

== Source-code and Models ==

Source-code and models available: [https://github.com/rodrigoberriel/satellite-crosswalk-classification GitHub].

== BibTeX ==

@article{Berriel2017cag,
author = {Rodrigo Ferreira Berriel and Franco Schmidt Rossi and Alberto Ferreira de Souza and Thiago Oliveira-Santos},
journal = {Computers & Graphics},
issn = {0097-8493}
title = {{Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach}},
volume = {68}
year = {2017},
month = {Nov},
pages = {32-42},
doi = {10.1016/J.CAG.2017.08.004}
}

Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach

2017-10-10T16:24:45Z

Rodrigo Berriel: /* BibTeX */

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], Franco Schmidt Rossi, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

DOI: [http://dx.doi.org/10.1016/J.CAG.2017.08.004 10.1016/J.CAG.2017.08.004]

PDF: [https://www.researchgate.net/publication/319194310_Automatic_Large-Scale_Data_Acquisition_via_Crowdsourcing_for_Crosswalk_Classification_A_Deep_Learning_Approach]

''Published in [https://www.journals.elsevier.com/computers-and-graphics/ Computers & Graphics]''

== Abstract ==

[[Arquivo:Crosswalk-streetview-graphical-abstract.png|x350px]]

Correctly identifying crosswalks is an essential task for the driving activity and mobility autonomy. Many crosswalk classification, detection and localization systems have been proposed in the literature over the years. These systems use different perspectives to tackle the crosswalk classification problem: satellite imagery, cockpit view (from the top of a car or behind the windshield), and pedestrian perspective. Most of the works in the literature are designed and evaluated using small and local datasets, i.e. datasets that present low diversity. Scaling to large datasets imposes a challenge for the annotation procedure. Moreover, there is still need for cross-database experiments in the literature because it is usually hard to collect the data in the same place and conditions of the final application. In this paper, we present a crosswalk classification system based on deep learning. For that, crowdsourcing platforms, such as OpenStreetMap and Google Street View, are exploited to enable automatic training via automatic acquisition and annotation of a large-scale database. Additionally, this work proposes a comparison study of models trained using fully-automatic data acquisition and annotation against models that were partially annotated. Cross-database experiments were also included in the experimentation to show that the proposed methods enable use with real world applications. Our results show that the model trained on the fully-automatic database achieved high overall accuracy (94.12%), and that a statistically significant improvement (to 96.30%) can be achieved by manually annotating a specific part of the database. Finally, the results of the cross-database experiments show that both models are robust to the many variations of image and scenarios, presenting a consistent behavior.

== Videos ==

See the IARA, GOPRO, and NIGHT dataset videos [https://www.youtube.com/playlist?list=PLm8amuguiXiIbZNQHj0BH1AJddnnql356 here].

== Source-code and Models ==

Source-code and models available: [https://github.com/rodrigoberriel/satellite-crosswalk-classification GitHub].

== BibTeX ==

@ARTICLE{Berriel2017cag,
author = {Rodrigo Ferreira Berriel and Franco Schmidt Rossi and Alberto Ferreira de Souza and Thiago Oliveira-Santos},
journal = {Computers & Graphics},
issn = {0097-8493}
title = {Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach},
volume = {68}
year = {2017},
month = {Nov},
pages = {32-42},
doi = {10.1016/J.CAG.2017.08.004}
}

Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach

2017-10-10T16:24:28Z

Rodrigo Berriel:

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], Franco Schmidt Rossi, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

DOI: [http://dx.doi.org/10.1016/J.CAG.2017.08.004 10.1016/J.CAG.2017.08.004]

PDF: [https://www.researchgate.net/publication/319194310_Automatic_Large-Scale_Data_Acquisition_via_Crowdsourcing_for_Crosswalk_Classification_A_Deep_Learning_Approach]

''Published in [https://www.journals.elsevier.com/computers-and-graphics/ Computers & Graphics]''

== Abstract ==

[[Arquivo:Crosswalk-streetview-graphical-abstract.png|x350px]]

Correctly identifying crosswalks is an essential task for the driving activity and mobility autonomy. Many crosswalk classification, detection and localization systems have been proposed in the literature over the years. These systems use different perspectives to tackle the crosswalk classification problem: satellite imagery, cockpit view (from the top of a car or behind the windshield), and pedestrian perspective. Most of the works in the literature are designed and evaluated using small and local datasets, i.e. datasets that present low diversity. Scaling to large datasets imposes a challenge for the annotation procedure. Moreover, there is still need for cross-database experiments in the literature because it is usually hard to collect the data in the same place and conditions of the final application. In this paper, we present a crosswalk classification system based on deep learning. For that, crowdsourcing platforms, such as OpenStreetMap and Google Street View, are exploited to enable automatic training via automatic acquisition and annotation of a large-scale database. Additionally, this work proposes a comparison study of models trained using fully-automatic data acquisition and annotation against models that were partially annotated. Cross-database experiments were also included in the experimentation to show that the proposed methods enable use with real world applications. Our results show that the model trained on the fully-automatic database achieved high overall accuracy (94.12%), and that a statistically significant improvement (to 96.30%) can be achieved by manually annotating a specific part of the database. Finally, the results of the cross-database experiments show that both models are robust to the many variations of image and scenarios, presenting a consistent behavior.

== Videos ==

See the IARA, GOPRO, and NIGHT dataset videos [https://www.youtube.com/playlist?list=PLm8amuguiXiIbZNQHj0BH1AJddnnql356 here].

== Source-code and Models ==

Source-code and models available: [https://github.com/rodrigoberriel/satellite-crosswalk-classification GitHub].

== BibTeX ==

@INPROCEEDINGS{Berriel2017cag,
author = {Rodrigo Ferreira Berriel and Franco Schmidt Rossi and Alberto Ferreira de Souza and Thiago Oliveira-Santos},
journal = {Computers & Graphics},
issn = {0097-8493}
title = {Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach},
volume = {68}
year = {2017},
month = {Nov},
pages = {32-42},
doi = {10.1016/J.CAG.2017.08.004}
}

Publicações do LCAD

2017-10-10T15:16:41Z

Rodrigo Berriel:

Ego-Lane Analysis System

2017-09-25T17:35:24Z

Rodrigo Berriel: add links

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo Berriel], [http://people.mpi-inf.mpg.de/~edeaguia/ Edilson de Aguiar], [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

'''Image and Vision Computing''': [https://doi.org/10.1016/j.imavis.2017.07.005 10.1016/j.imavis.2017.07.005]

== Abstract ==

[[Arquivo:Elas-graphical-abstract.png|x350px]]

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines, Kalman filter and Particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e. lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.

== Demonstration Video ==

Demonstration Video of ELAS, as proposed by the time of submission

[[Arquivo:Elas-video1.png|x200px|link=http://bit.ly/SI-AUTO-VISION_DEMO-VIDEO]]

ELAS was weakly integrated into IARA (our autonomous vehicle). The video below shows ELAS performing on IARA (without tuning any parameter).

[[Arquivo:Elas-carmen.png|x200px|link=https://www.youtube.com/watch?v=R5wdPJ4ZI5M]]

== Source-Code ==
Available [https://github.com/rodrigoberriel/ego-lane-analysis-system here]

== Dataset ==
To request access to the datasets, read the instructions [https://github.com/rodrigoberriel/ego-lane-analysis-system/blob/master/datasets/ here].

It contains 22 scenes with a total of 17,092 frames. In 98.54% of the frames, at least one of the sides has white lane markings, while 12.88% of them are yellow. In 5.18% of the frames at least one side has no lane markings. This numbers include frames containing transitions of lane markings type, i.e. images with two lane markings type simultaneously on the same side. Lane markings type transitions are present in 1,854 frames (10.85%). Lane change maneuver are being performed in 5.42% of this dataset, where 64.72% is from right to left and 35.28% in the opposite direction. There are intersections in 2.11% of the frames. In 7.28% of all images, there is at least one pavement marking. 33.92% of the pavement markings are crosswalks and 15.27% are annotated as unknown, i.e. they are of none of the classes of interest and 50.81% comprises arrows and stop lines. There is at least one adjacent lane in 50.34% and 72.14% of the frames for each side, right and left respectively.

[http://www.lcad.inf.ufes.br/~berriel/paper-si-auto-vision/dataset-short-description.pdf Download a short description of each scene (with images).]

Some samples below:

[[Arquivo:Elas-dataset-samples.png|x300px]]

== BibTeX ==

@article{Berriel2017imavis,
author = "Rodrigo F. Berriel and Edilson de Aguiar and Alberto F. de Souza and Thiago Oliveira-Santos",
title = "Ego-Lane Analysis System (ELAS): Dataset and Algorithms",
journal = "Image and Vision Computing",
year = "2017",
note = "In Press",
issn = "0262-8856",
doi = "https://doi.org/10.1016/j.imavis.2017.07.005",
}

Ego-Lane Analysis System

2017-08-13T13:31:57Z

Rodrigo Berriel: doi and bibtex

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo Berriel], [http://people.mpi-inf.mpg.de/~edeaguia/ Edilson de Aguiar], [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

'''Image and Vision Computing''': [https://doi.org/10.1016/j.imavis.2017.07.005 10.1016/j.imavis.2017.07.005]

== Abstract ==

[[Arquivo:Elas-graphical-abstract.png|x350px]]

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines, Kalman filter and Particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e. lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.

== Demonstration Video ==

Demonstration Video of ELAS, as proposed by the time of submission

[[Arquivo:Elas-video1.png|x200px|link=http://bit.ly/SI-AUTO-VISION_DEMO-VIDEO]]

ELAS was weakly integrated into IARA (our autonomous vehicle). The video below shows ELAS performing on IARA (without tuning any parameter).

[[Arquivo:Elas-carmen.png|x200px|link=https://www.youtube.com/watch?v=R5wdPJ4ZI5M]]

== Source-Code ==
Available upon acceptance.

== Dataset ==
Available upon acceptance.

It contains 22 scenes with a total of 17,092 frames. In 98.54% of the frames, at least one of the sides has white lane markings, while 12.88% of them are yellow. In 5.18% of the frames at least one side has no lane markings. This numbers include frames containing transitions of lane markings type, i.e. images with two lane markings type simultaneously on the same side. Lane markings type transitions are present in 1,854 frames (10.85%). Lane change maneuver are being performed in 5.42% of this dataset, where 64.72% is from right to left and 35.28% in the opposite direction. There are intersections in 2.11% of the frames. In 7.28% of all images, there is at least one pavement marking. 33.92% of the pavement markings are crosswalks and 15.27% are annotated as unknown, i.e. they are of none of the classes of interest and 50.81% comprises arrows and stop lines. There is at least one adjacent lane in 50.34% and 72.14% of the frames for each side, right and left respectively.

[http://www.lcad.inf.ufes.br/~berriel/paper-si-auto-vision/dataset-short-description.pdf Download a short description of each scene (with images).]

Some samples below:

[[Arquivo:Elas-dataset-samples.png|x300px]]

== BibTeX ==

@article{Berriel2017imavis,
author = "Rodrigo F. Berriel and Edilson de Aguiar and Alberto F. de Souza and Thiago Oliveira-Santos",
title = "Ego-Lane Analysis System (ELAS): Dataset and Algorithms",
journal = "Image and Vision Computing",
year = "2017",
note = "In Press",
issn = "0262-8856",
doi = "https://doi.org/10.1016/j.imavis.2017.07.005",
}

Cross-Database Facial Expression Recognition Based on Fine-Tuned Deep Convolutional Network

2017-08-04T18:59:28Z

Rodrigo Berriel: init

More information soon.

Abstract: Facial expression recognition is a very important research field to understand human emotions. Many facial expression recognition systems have been proposed in the literature over the years. Some of these methods use neural network approaches with deep architectures to address the problem. Although it seems that the facial expression recognition problem has been solved, there is a large difference between the results achieved using the same database to train and test the network and the cross-database protocol. In this paper, we extensively investigate the performance influence of fine-tuning with crossdatabase approach. In order to perform the study, the VGGFace Deep Convolutional Network model (pre-trained for face recognition) was fine-tuned to recognize facial expressions considering different well-established databases in the literature: CK+, JAFFE, MMI, RaFD, KDEF, BU3DFE, and AR Face. The cross-database experiments were organized so that one of the databases was separated as test set and the others as training,
and each experiment was ran multiple times to ensure the results. Our results show a significant improvement on the use of pre-trained models against randomly initialized Convolutional Neural Networks on the facial expression recognition problem, for example achieving 88.58%, 67.03%, 85.97%, and 72.55% average accuracy testing in the CK+, MMI, RaFD, and KDEF, respectively. Additionally, in absolute terms, the results show an improvement in the literature for cross-database facial expression recognition with the use of pre-trained models.

Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach

2017-07-25T21:36:34Z

Rodrigo Berriel: add videos

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], Franco Schmidt Rossi, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

''To appear in [https://www.journals.elsevier.com/computers-and-graphics/ Computers & Graphics]''

== Abstract ==

[[Arquivo:Crosswalk-streetview-graphical-abstract.png|x350px]]

Correctly identifying crosswalks is an essential task for the driving activity and mobility autonomy. Many crosswalk classification, detection and localization systems have been proposed in the literature over the years. These systems use different perspectives to tackle the crosswalk classification problem: satellite imagery, cockpit view (from the top of a car or behind the windshield), and pedestrian perspective. Most of the works in the literature are designed and evaluated using small and local datasets, i.e. datasets that present low diversity. Scaling to large datasets imposes a challenge for the annotation procedure. Moreover, there is still need for cross-database experiments in the literature because it is usually hard to collect the data in the same place and conditions of the final application. In this paper, we present a crosswalk classification system based on deep learning. For that, crowdsourcing platforms, such as OpenStreetMap and Google Street View, are exploited to enable automatic training via automatic acquisition and annotation of a large-scale database. Additionally, this work proposes a comparison study of models trained using fully-automatic data acquisition and annotation against models that were partially annotated. Cross-database experiments were also included in the experimentation to show that the proposed methods enable use with real world applications. Our results show that the model trained on the fully-automatic database achieved high overall accuracy (94.12%), and that a statistically significant improvement (to 96.30%) can be achieved by manually annotating a specific part of the database. Finally, the results of the cross-database experiments show that both models are robust to the many variations of image and scenarios, presenting a consistent behavior.

== Videos ==

See the IARA, GOPRO, and NIGHT dataset videos [https://www.youtube.com/playlist?list=PLm8amuguiXiIbZNQHj0BH1AJddnnql356 here].

== Available Soon ==

Source-code, and other resources.

Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach

2017-07-25T21:32:44Z

Rodrigo Berriel: correção no nome do journal

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], Franco Schmidt Rossi, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

''To appear in [https://www.journals.elsevier.com/computers-and-graphics/ Computers & Graphics]''

== Abstract ==

[[Arquivo:Crosswalk-streetview-graphical-abstract.png|x350px]]

Correctly identifying crosswalks is an essential task for the driving activity and mobility autonomy. Many crosswalk classification, detection and localization systems have been proposed in the literature over the years. These systems use different perspectives to tackle the crosswalk classification problem: satellite imagery, cockpit view (from the top of a car or behind the windshield), and pedestrian perspective. Most of the works in the literature are designed and evaluated using small and local datasets, i.e. datasets that present low diversity. Scaling to large datasets imposes a challenge for the annotation procedure. Moreover, there is still need for cross-database experiments in the literature because it is usually hard to collect the data in the same place and conditions of the final application. In this paper, we present a crosswalk classification system based on deep learning. For that, crowdsourcing platforms, such as OpenStreetMap and Google Street View, are exploited to enable automatic training via automatic acquisition and annotation of a large-scale database. Additionally, this work proposes a comparison study of models trained using fully-automatic data acquisition and annotation against models that were partially annotated. Cross-database experiments were also included in the experimentation to show that the proposed methods enable use with real world applications. Our results show that the model trained on the fully-automatic database achieved high overall accuracy (94.12%), and that a statistically significant improvement (to 96.30%) can be achieved by manually annotating a specific part of the database. Finally, the results of the cross-database experiments show that both models are robust to the many variations of image and scenarios, presenting a consistent behavior.

== Available Soon ==

Videos, source-code, and other resources.

Automatic Large-Scale Data Acquisition via Crowdsourcing for Crosswalk Classification: A Deep Learning Approach

2017-07-24T18:16:41Z

Rodrigo Berriel: Criou página com 'category:Publicações Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], Franco Schmidt Rossi, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferrei...'

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo F. Berriel], Franco Schmidt Rossi, [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

''To appear in [https://www.journals.elsevier.com/computers-and-graphics/ Computer & Graphics]''

== Abstract ==

[[Arquivo:Crosswalk-streetview-graphical-abstract.png|x350px]]

Correctly identifying crosswalks is an essential task for the driving activity and mobility autonomy. Many crosswalk classification, detection and localization systems have been proposed in the literature over the years. These systems use different perspectives to tackle the crosswalk classification problem: satellite imagery, cockpit view (from the top of a car or behind the windshield), and pedestrian perspective. Most of the works in the literature are designed and evaluated using small and local datasets, i.e. datasets that present low diversity. Scaling to large datasets imposes a challenge for the annotation procedure. Moreover, there is still need for cross-database experiments in the literature because it is usually hard to collect the data in the same place and conditions of the final application. In this paper, we present a crosswalk classification system based on deep learning. For that, crowdsourcing platforms, such as OpenStreetMap and Google Street View, are exploited to enable automatic training via automatic acquisition and annotation of a large-scale database. Additionally, this work proposes a comparison study of models trained using fully-automatic data acquisition and annotation against models that were partially annotated. Cross-database experiments were also included in the experimentation to show that the proposed methods enable use with real world applications. Our results show that the model trained on the fully-automatic database achieved high overall accuracy (94.12%), and that a statistically significant improvement (to 96.30%) can be achieved by manually annotating a specific part of the database. Finally, the results of the cross-database experiments show that both models are robust to the many variations of image and scenarios, presenting a consistent behavior.

== Available Soon ==

Videos, source-code, and other resources.

Arquivo:Crosswalk-streetview-graphical-abstract.png

2017-07-24T18:13:09Z

Rodrigo Berriel:

Equipe

2017-03-25T16:54:11Z

Rodrigo Berriel:

__TOC__

==Researchers==

* [http://www.lcad.inf.ufes.br/team/index.php/Dr._Alberto_Ferreira_De_Souza Dr. Alberto Ferreira De Souza (Coordinator)]
* Dr. Andrea Maria Pedrosa Valli
* Dr. Claudine Badue
* Dr. Edilson de Aguiar
* Dr. Elias Oliveira
* [http://www.lcad.inf.ufes.br/team/index.php/Dr._Fabio_Daros_Freitas Dr. Fabio Daros Freitas]
* Dr. Lucia Catabriga
* Dr. Maria Claudia Silva Boeres
* Dr. Maria Cristina Rangel
* Dr. Thiago Oliveira dos Santos

==Phd Students==

* [http://www.lcad.inf.ufes.br/~avelino Avelino Forechi]
* [http://www.lcad.inf.ufes.br/~lveronese Lucas de Paula Veronese]
* [http://www.inf.ufes.br/~mberger Mariella Berger]
* [http://rodrigoberriel.com Rodrigo Berriel]

==Masters Students==

* Cayo Fontana
* [http://www.lcad.inf.ufes.br/~filipe Filipe Wall Mutz]
* Lauro Jose Lyrio Junior
* Michael André Goncalves
* Romulo Ramos Radaelli
* [http://www.lcad.inf.ufes.br/~toliveira Tiago Alves de Oliveira]
* Vitor Barbirato

==Graduate Students==

* Lucas Catabriga
* Rafael Correia Nascimento
* Leornado Cunha
* Ranick Guidolini

Ego-Lane Analysis System

2017-03-25T16:45:47Z

Rodrigo Berriel: add na categoria

[[category:Publicações]]
Authors: [http://rodrigoberriel.com/ Rodrigo Berriel], [http://people.mpi-inf.mpg.de/~edeaguia/ Edilson de Aguiar], [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

''This paper is under review (Special Issue: Automotive Vision, Image and Vision Computing)''

== Abstract ==

[[Arquivo:Elas-graphical-abstract.png|x350px]]

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines, Kalman filter and Particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e. lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.

== Demonstration Video ==

Demonstration Video of ELAS, as proposed by the time of submission

[[Arquivo:Elas-video1.png|x200px|link=http://bit.ly/SI-AUTO-VISION_DEMO-VIDEO]]

ELAS was weakly integrated into IARA (our autonomous vehicle). The video below shows ELAS performing on IARA (without tuning any parameter).

[[Arquivo:Elas-carmen.png|x200px|link=https://www.youtube.com/watch?v=R5wdPJ4ZI5M]]

== Source-Code ==
Available upon acceptance.

== Dataset ==
Available upon acceptance.

It contains 22 scenes with a total of 17,092 frames. In 98.54% of the frames, at least one of the sides has white lane markings, while 12.88% of them are yellow. In 5.18% of the frames at least one side has no lane markings. This numbers include frames containing transitions of lane markings type, i.e. images with two lane markings type simultaneously on the same side. Lane markings type transitions are present in 1,854 frames (10.85%). Lane change maneuver are being performed in 5.42% of this dataset, where 64.72% is from right to left and 35.28% in the opposite direction. There are intersections in 2.11% of the frames. In 7.28% of all images, there is at least one pavement marking. 33.92% of the pavement markings are crosswalks and 15.27% are annotated as unknown, i.e. they are of none of the classes of interest and 50.81% comprises arrows and stop lines. There is at least one adjacent lane in 50.34% and 72.14% of the frames for each side, right and left respectively.

[http://www.lcad.inf.ufes.br/~berriel/paper-si-auto-vision/dataset-short-description.pdf Download a short description of each scene (with images).]

Some samples below:

[[Arquivo:Elas-dataset-samples.png|x300px]]

Ego-Lane Analysis System

2017-02-03T17:46:03Z

Rodrigo Berriel: adição de imagens

Authors: [http://rodrigoberriel.com/ Rodrigo Berriel], [http://people.mpi-inf.mpg.de/~edeaguia/ Edilson de Aguiar], [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

''This paper is under review (Special Issue: Automotive Vision, Image and Vision Computing)''

== Abstract ==

[[Arquivo:Elas-graphical-abstract.png|x350px]]

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines, Kalman filter and Particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e. lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.

== Demonstration Video ==

Demonstration Video of ELAS, as proposed by the time of submission

[[Arquivo:Elas-video1.png|x200px|link=http://bit.ly/SI-AUTO-VISION_DEMO-VIDEO]]

ELAS was weakly integrated into IARA (our autonomous vehicle). The video below shows ELAS performing on IARA (without tuning any parameter).

[[Arquivo:Elas-carmen.png|x200px|link=https://www.youtube.com/watch?v=R5wdPJ4ZI5M]]

== Source-Code ==
Available upon acceptance.

== Dataset ==
Available upon acceptance.

It contains 22 scenes with a total of 17,092 frames. In 98.54% of the frames, at least one of the sides has white lane markings, while 12.88% of them are yellow. In 5.18% of the frames at least one side has no lane markings. This numbers include frames containing transitions of lane markings type, i.e. images with two lane markings type simultaneously on the same side. Lane markings type transitions are present in 1,854 frames (10.85%). Lane change maneuver are being performed in 5.42% of this dataset, where 64.72% is from right to left and 35.28% in the opposite direction. There are intersections in 2.11% of the frames. In 7.28% of all images, there is at least one pavement marking. 33.92% of the pavement markings are crosswalks and 15.27% are annotated as unknown, i.e. they are of none of the classes of interest and 50.81% comprises arrows and stop lines. There is at least one adjacent lane in 50.34% and 72.14% of the frames for each side, right and left respectively.

[http://www.lcad.inf.ufes.br/~berriel/paper-si-auto-vision/dataset-short-description.pdf Download a short description of each scene (with images).]

Some samples below:

[[Arquivo:Elas-dataset-samples.png|x300px]]

Arquivo:Elas-dataset-samples.png

2017-02-03T17:44:14Z

Rodrigo Berriel:

Arquivo:Elas-carmen.png

2017-02-03T17:38:12Z

Rodrigo Berriel:

Arquivo:Elas-video1.png

2017-02-03T17:37:58Z

Rodrigo Berriel:

Arquivo:Elas-graphical-abstract.png

2017-02-03T17:29:12Z

Rodrigo Berriel:

Ego-Lane Analysis System

2016-09-05T14:17:47Z

Rodrigo Berriel: /* Dataset */

Authors: [http://rodrigoberriel.com/ Rodrigo Berriel], [http://people.mpi-inf.mpg.de/~edeaguia/ Edilson de Aguiar], [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

''This paper is under review (Special Issue: Automotive Vision, Image and Vision Computing)''

[http://www.lcad.inf.ufes.br/~berriel/paper-si-auto-vision/images/graphical-abstract.png Graphical Abstract]

== Abstract ==

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines, Kalman filter and Particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e. lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.

== Demonstration Video ==
[http://bit.ly/SI-AUTO-VISION_DEMO-VIDEO Demonstration Video of ELAS]

== Dataset ==
Available upon acceptance.

It contains 22 scenes with a total of 17,092 frames. In 98.54% of the frames, at least one of the sides has white lane markings, while 12.88% of them are yellow. In 5.18% of the frames at least one side has no lane markings. This numbers include frames containing transitions of lane markings type, i.e. images with two lane markings type simultaneously on the same side. Lane markings type transitions are present in 1,854 frames (10.85%). Lane change maneuver are being performed in 5.42% of this dataset, where 64.72% is from right to left and 35.28% in the opposite direction. There are intersections in 2.11% of the frames. In 7.28% of all images, there is at least one pavement marking. 33.92% of the pavement markings are crosswalks and 15.27% are annotated as unknown, i.e. they are of none of the classes of interest and 50.81% comprises arrows and stop lines. There is at least one adjacent lane in 50.34% and 72.14% of the frames for each side, right and left respectively.

[http://www.lcad.inf.ufes.br/~berriel/paper-si-auto-vision/dataset-short-description.pdf Short description of each scene.]

== Source-Code ==
Available upon acceptance.

Ego-Lane Analysis System

2016-09-05T14:10:05Z

Rodrigo Berriel:

Authors: [http://rodrigoberriel.com/ Rodrigo Berriel], [http://people.mpi-inf.mpg.de/~edeaguia/ Edilson de Aguiar], [http://www.lcad.inf.ufes.br/team/index.php/Prof._Dr._Alberto_Ferreira_De_Souza Alberto F. de Souza], [http://www.inf.ufes.br/~todsantos/home Thiago Oliveira-Santos]

''This paper is under review (Special Issue: Automotive Vision, Image and Vision Computing)''

[http://www.lcad.inf.ufes.br/~berriel/paper-si-auto-vision/images/graphical-abstract.png Graphical Abstract]

== Abstract ==

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines, Kalman filter and Particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e. lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.

== Demonstration Video ==
[http://bit.ly/SI-AUTO-VISION_DEMO-VIDEO Demonstration Video of ELAS]

== Dataset ==
Available upon acceptance.

It contains 22 scenes with a total of 17,092 frames. In 98.54% of the frames, at least one of the sides has white lane markings, while 12.88% of them are yellow. In 5.18% of the frames at least one side has no lane markings. This numbers include frames containing transitions of lane markings type, i.e. images with two lane markings type simultaneously. Lane markings type transitions are present in 1,854 frames (10.85%). Lane change maneuver are being performed in 5.42% of this dataset, where 64.72% is from right to left and 35.28% in the opposite direction. There are intersections in 2.11% of the frames. In 7.28% of all images, there is at least one pavement marking. 33.92% of the pavement markings are crosswalks and 15.27% are annotated as unknown, i.e. they are of none of the classes of interest and 50.81% comprises arrows and stop lines. There is at least one adjacent lane in 50.34% and 72.14% of the frames for each side, right and left respectively.

[http://www.lcad.inf.ufes.br/~berriel/paper-si-auto-vision/dataset-short-description.pdf Short description of each scene.]

== Source-Code ==
Available upon acceptance.

Ego-Lane Analysis System

2016-08-09T12:32:14Z

Rodrigo Berriel:

Ego-Lane Analysis System

2016-08-09T12:30:59Z

Rodrigo Berriel: adicionei ponteiro para o outro vídeo e outro para uma imagem da IARA

A Particle Filter-based Lane Marker Tracking Approach using a Cubic Spline Model

2016-07-21T12:46:26Z

Rodrigo Berriel: add link para o pdf

[[category:Publicações]]
Conference Paper, SIBGRAPI, 2015

DOI: [http://dx.doi.org/10.1109/SIBGRAPI.2015.15 10.1109/SIBGRAPI.2015.15]

PDF: [http://sibgrapi.sid.inpe.br/col/sid.inpe.br/sibgrapi/2015/06.19.21.39/doc/PID3755347.pdf]

== Authors ==
Rodrigo Berriel, Edilson de Aguiar, Vanderlei Vieira de Souza Filho, Thiago Oliveira-Santos

== Abstract ==
In this paper we present a particle filter-based lane marker tracking approach using a cubic spline model. The system can detect the two main lane markers (i.e. lane strips) of marked roads from a monocular camera mounted on the top of a vehicle. Traditional lane marker detection and tracking systems have limitations to properly detect curved roads and to use temporal information to better estimate and track lane markers. The proposed system works on a temporal sequence of images. For each image, one at time, it applies a sequence of steps comprising an inverse perspective mapping to correct for perspective distortions, and a particle filter to smoothly track the lane markers along time. The output of the system is a lane marker generated by a cubic spline interpolation scheme to fit a wider range of lanes. Our system can run in real applications and it was validated with various road and traffic conditions. As a result, it achieves a high precision (98.13%) and a small error (0.0143 meters).

[[imagem:Resumo_A_Particle_Filter-based_Lane_Marker_Tracking_Approach_using_a_Cubic_Spline_Model.jpg]]

== Video ==
[https://youtu.be/cI79RZhIDCI Demo]

== BibTeX ==

@INPROCEEDINGS{Berriel2015,
author = {Rodrigo Ferreira Berriel and Edilson de Aguiar and Vanderlei Vieira de Souza Filho and Thiago Oliveira-Santos},
booktitle = {Graphics, Patterns and Images (SIBGRAPI), 2015 28th SIBGRAPI Conference on},
title = {A Particle Filter-Based Lane Marker Tracking Approach Using a Cubic Spline Model},
year = {2015},
month = {Aug},
pages = {149-156},
doi = {10.1109/SIBGRAPI.2015.15}
}

Instalação Carmen para Ubuntu 12.04.3

2016-04-15T19:00:43Z

Rodrigo Berriel: precisa de sudo para os ln

Atualizar o apt:
sudo apt-get update

Atualizar os pacotes:
sudo apt-get dist-upgrade

Instalar o subversion, git, gimp, meld e vim:
sudo apt-get install gimp meld subversion vim git

Baixar o Carmen pelo git (faça o download enquanto segue os próximos passos):
git clone https://github.com/LCAD-UFES/carmen_lcad.git

Baixe a MAE via git:
git clone https://github.com/LCAD-UFES/MAE.git

Instalar os pacotes para o carmem:
sudo apt-get install swig \
libgtk2.0-dev \
qt-sdk \
libqt3-mt libqt3-mt-dev qt3-dev-tools \
libimlib2 libimlib2-dev \
imagemagick libmagick++-dev \
libwrap0 libwrap0-dev tcpd \
libncurses5 libncurses5-dev \
libgsl0-dev libgsl0ldbl \
libdc1394-22 libdc1394-22-dev libdc1394-utils \
cmake \
libgtkglext1 libgtkglext1-dev \
libgtkglextmm-x11-1.2-0 libgtkglextmm-x11-1.2-dev \
libglade2-0 libglade2-dev \
freeglut3 freeglut3-dev \
libcurl3 libcurl3-nss libcurl4-nss-dev \
libglew1.5 libglew1.5-dev libglewmx1.5 libglewmx1.5-dev glew-utils1.5 \
libkml0 libkml-dev \
liburiparser1 liburiparser-dev \
git \
libusb-1.0-0 libusb-1.0-0-dev libusb-dev \
libxi-dev libxi6 \
libxmu-dev libxmu6 \
build-essential libforms-dev \
byacc \
flex \
doxygen \
libespeak-dev

caso de erro em libcheese-gtk23 instale os pacotes abaixo antes dos anteriores (Solucão nao confirmada):
sudo apt-get install libglew-dev libcheese7 libcheese-gtk23 libclutter-gst-2.0-0 libcogl15 libclutter-gtk-1.0-0 libclutter-1.0-0

Instalar o Java:
sudo add-apt-repository ppa:webupd8team/java
sudo apt-get update && sudo apt-get install oracle-jdk7-installer
update-alternatives --display java

Edite o arquivo /etc/environment:
sudo gedit /etc/environment

Adicione no final do arquivo:
JAVA_HOME=/usr/lib/jvm/java-7-oracle

Instalar o eclipse:

Baixe o eclipse de:
http://www.eclipse.org/downloads/

Descompacte o ecplise
cd Downloads/
sudo mv eclipse-cpp-kepler-R-linux-gtk-x86_64.tar.gz /opt
cd /opt/
sudo tar -xvf eclipse-cpp-kepler-R-linux-gtk-x86_64.tar.gz

Crie um arquivo desktop e edite ele em /usr/share/applications:
sudo gedit /usr/share/applications/eclipse.desktop

Coloque o seguinte conteudo:
[Desktop Entry]
Name=Eclipse
Type=Application
Exec=/opt/eclipse/eclipse
Terminal=false
Icon=/opt/eclipse/icon.xpm
Comment=Integrated Development Environment
NoDisplay=false
Categories=Development;IDE
Name[en]=Eclipse

Instale as deps da PCL:
sudo apt-get install libeigen3-dev libboost-all-dev libflann-dev libvtk5-dev cmake-gui

Instale as deps do OpenCV:
sudo apt-get install build-essential libgtk2.0-dev libjpeg-dev libtiff4-dev libjasper-dev \
libopenexr-dev cmake python-dev python-numpy python-tk libtbb-dev libeigen2-dev yasm libfaac-dev \
libopencore-amrnb-dev libopencore-amrwb-dev libtheora-dev libvorbis-dev libxvidcore-dev libx264-dev \
libqt4-dev libqt4-opengl-dev sphinx-common texlive-latex-extra libv4l-dev libdc1394-22-dev \
libavcodec-dev libavformat-dev libswscale-dev

Baixar os arquivos:
sudo su
cd /usr/local/
wget http://bullet.googlecode.com/files/bullet-2.78-r2387.tgz
wget http://downloads.sourceforge.net/project/opencvlibrary/opencv-unix/2.4.9/opencv-2.4.9.zip
wget http://downloads.sourceforge.net/project/fann/fann/2.2.0/FANN-2.2.0-Source.tar.gz
wget http://www.kvaser.com/software/7330130980754/V5_3_0/linuxcan.tar.gz
tar -xvf bullet-2.78-r2387.tgz
unzip opencv-2.4.9.zip
tar -xvf linuxcan.tar.gz
tar -xvf FANN-2.2.0-Source.tar.gz
mv bullet-2.78 bullet
cd bullet
./configure
make
make install
cd ..
cd linuxcan
make
make install
cd ..
cd FANN-2.2.0-Source
mkdir build
cd build
cmake ..
make
make install
cd ../..
cd opencv-2.4.9
mkdir build
cd build
cmake -D WITH_TBB=ON -D WITH_CUDA=OFF -D BUILD_NEW_PYTHON_SUPPORT=ON -D WITH_V4L=ON -D INSTALL_C_EXAMPLES=ON -D INSTALL_PYTHON_EXAMPLES=ON -D BUILD_EXAMPLES=ON -D WITH_QT=ON -D WITH_OPENGL=ON ..
make
make install

Edite o arquivo /etc/ld.so.conf.d/opencv.conf
gedit /etc/ld.so.conf.d/opencv.conf

Adicione ao final dele:
/usr/local/lib

Execute:
ldconfig

Edite o arquivo /etc/bash.bashrc:
gedit /etc/bash.bashrc

Adicione no final do arquivo:
PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig
export PKG_CONFIG_PATH

sair do rooot
exit

Coloque no .bashrc:
#CARMEN
export PKG_CONFIG_PATH=$PKG_CONFIG_PATH:/usr/local/lib/pkgconfig
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/lib:/usr/lib/x86_64-linux-gnu:/usr/lib/i386-linux-gnu/:/usr/lib/libkml
export CARMEN_HOME=~/carmen_lcad

#MAE
export MAEHOME=~/MAE
export PATH=$PATH:$MAEHOME/bin

Instale os pacotes imlib e flycapture:
cd $CARMEN_HOME/ubuntu_packages/
sudo dpkg -i imlib_1.9.15-20_amd64.deb
sudo dpkg -i imlib-devel_1.9.15-20_amd64.deb
tar -xvf flycapture2-2.5.3.4-amd64-pkg.tgz
cd flycapture2-2.5.3.4-amd64/
sudo apt-get install libglademm-2.4-1c2a
sudo apt-get install libglademm-2.4-dev
sudo apt-get install libgtkmm-2.4-dev
sudo sh install_flycapture.sh

Faça os links:
sudo ln -s /usr/lib64/libgdk_imlib.so.1.9.15 /usr/lib64/libgdk_imlib.a
sudo ln -s /usr/src/linux-headers-3.8.0-30/ /usr/src/linux

Instale a PCL:
sudo add-apt-repository ppa:v-launchpad-jochen-sprickerhof-de/pcl
sudo apt-get update
sudo apt-get install libpcl-all

Instale a câmera Kinect:
sudo su
cd /usr/local
wget http://sourceforge.net/projects/libusb/files/libusb-1.0/libusb-1.0.19/libusb-1.0.19.tar.bz2
tar xvf libusb-1.0.19.tar.bz2
cd libusb-1.0.19
./configure
make
make install

se der erro na instalacao acima tente instalar a udev-dev antes:
sudo apt-get install libudev-dev

mkdir /usr/local/tplib
cd /usr/local/tplib
git clone git://github.com/OpenKinect/libfreenect.git
cd libfreenect
mkdir build
cd build
cmake ..
cp src/libfreenect.pc /usr/local/tplib/
make
cp ../src/libfreenect.pc.in src/libfreenect.pc
cp ../fakenect/fakenect.sh.in fakenect/fakenect.sh
make install
ldconfig /usr/local/lib64/
exit

Execute:
glview

Caso dê erro, tente:
freenect-glview

Se der erro execute:
sudo ldconfig /usr/local/lib64/

Instalação da biblioteca G2O:
sudo apt-get install cmake libsuitesparse-dev libqt4-dev qt4-qmake libqglviewer-qt4-dev
cd /usr/local/
sudo svn co https://svn.openslam.org/data/svn/g2o
cd /usr/local/g2o/trunk/build/
sudo cmake ../ -DBUILD_CSPARSE=ON -DG2O_BUILD_DEPRECATED_TYPES=ON -DG2O_BUILD_LINKED_APPS=ON
sudo make
sudo make install

Alterações extras para o ubuntu 14.04 (não faça as mudanças caso você use o 12.04):
cd $CARMEN_HOME/ubuntu_packages/
sudo dpkg -i zlib1g_1.2.3.4.dfsg-3ubuntu4_amd64.deb
sudo dpkg -i zlib1g-dev_1.2.3.4.dfsg-3ubuntu4_amd64.deb
sudo ln -s /usr/lib/x86_64-linux-gnu/libboost_thread.so /usr/lib/x86_64-linux-gnu/libboost_thread-mt.so

Observações importantes:
Os pacotes libqt3-mt libqt3-mt-dev qt3-dev-tools do inicio da instalação não existem no 14.04, porém é possivel que ninguem use mais o QT3 no carmen, caso alguém saiba sobre o uso do QT3 fazer as devidas alterações
O pacote glew-utils1.5 não existe mais, porém instalei o pacote libglew-dev e o carmen compilou sem problemas. Favor verificar se essa mudança é valida e se ela também pode ser aplicada ao 12.04
O pacote libqglviewer-qt4-dev não tem candidatos no ubuntu 14.04, ele é realmente necessário? Caso alguem faça uma instalação limpa no ubuntu 12.04 verificar se fica tudo ok sem ele.

Instale as bibliotecas da MAE:
apt-get install make g++ freeglut3-dev byacc libforms-dev libtiff4-dev libXi-dev libXmu-dev doxygen tcsh flex libdc1394-22-dev

# Compilar a MAE:
## Pré-requisito: as variáveis de ambiente MAEHOME e PATH devem estar ajustadas;
## Entrar no diretório da MAE: "cd $MAEHOME";
## Compilar a MAE: "make".
## Verificar se a biblioteca da MAE libnet_conn.a foi gerado em MAEHOME/lib;
## Verificar se o compilador da MAE netcomp foi gerado em MAEHOME/bin.

Mais informações sobre a MAE:
http://www.lcad.inf.ufes.br/wiki/index.php/M%C3%A1quina_Associadora_de_Eventos_-_MAE#Compilando_a_MAE

Feche todos os terminais e faça:
cd $CARMEN_HOME/src
./configure --nojava --nocuda
Should the C++ tools be installed for CARMEN: [Y/n] Y
Should Python Bindings be installed: [y/N] y
Searching for Python2.4... Should the old laser server be used instead of the new one: [y/N] N
Install path [/usr/local/]:
Robot numbers [*]: 1,2

Antes de compilar o CARMEN, precisamos que o módulo tracker seja compilado separadamente para que o navigator_spline funcione:
cd $CARMEN_HOME/src/tracker
make

Vai dar um erro de compilação, mas está tudo ok.

Para compilar o carmen rode:
cd $CARMEN_HOME/src
make

Caso dê erro por causa da libusb.h vá no arquivo:
sudo vim /usr/local/include/libfreenect.hpp
E altere #include <libusb.h> para
#include <libusb-1.0/libusb.h>

== Ajustes na rede para o GPS Trimble ==

Para conectar o novo GPS Trimble é necessário uma conexão com a Internet dentro da IARA. Optamos por usar um iPhone com conexão 3G.

Para o iPhone funcionar no Ubuntu 12.04 é necessário um tanto de coisas... Perdemos o histórico mas dá para achar na Internet (Google iPhone 4S ubuntu 12.04 mount). Precisa instalar uns pacotes (apt-get install ...). Se você tiver sucesso, vai ser possível usar o iPhone como Personal Hotspot, ou seja, usar a Internet de dentro da IARA.

Feito isso, é necessário criar um Gateway da máquina que tem acesso a Internet (car01) para uma subrede da IARA (192.168.0.0 - a subrede de Carro Network). Para isso (ver página de referência em https://help.ubuntu.com/community/Internet/ConnectionSharing (Gateway set up)), considerando o iPhone em eth2 e a subrede da IARA em eth1:
sudo iptables -A FORWARD -o eth2 -i eth1 -s 192.168.0.0/24 -m conntrack --ctstate NEW -j ACCEPT
sudo iptables -A FORWARD -m conntrack --ctstate ESTABLISHED,RELATED -j ACCEPT
sudo iptables -t nat -F POSTROUTING
sudo iptables -t nat -A POSTROUTING -o eth2 -j MASQUERADE
sudo iptables-save | sudo tee /etc/iptables.sav

Os comandos acima criam um NAT do iPhone para a subrede da IARA. Em seguida, é necessário editar o /etc/rc.local e adicionar a linha abaixo antes de "exit 0":
iptables-restore < /etc/iptables.sav

É necessário ainda:
sudo sh -c "echo 1 > /proc/sys/net/ipv4/ip_forward"

Para tornar isso permanente, inclua as linhas abaixo em /etc/sysctl.conf:
net.ipv4.ip_forward=1
net.ipv4.conf.default.forwarding=1
net.ipv4.conf.all.forwarding=1

No Network Manager, tem que setar "Use this connection only for resources on its network" (Network Manager->IPv4 Settings->Routes) em todas as redes cabeadas exceto a do iPhone.

Pronto!

Ego-Lane Analysis System

2016-04-12T17:42:02Z

Rodrigo Berriel: troquei IP pelo endereço

Ego-Lane Analysis System

2016-04-11T17:57:51Z

Rodrigo Berriel: add link para graphical abstract

Ego-Lane Analysis System

2016-04-11T17:45:40Z

Rodrigo Berriel: remove para não ter como chegar ao link, durante review

Ego-Lane Analysis System

2016-04-06T12:07:14Z

Rodrigo Berriel: add link para rastrear acessos

Ego-Lane Analysis System

2016-04-04T17:26:20Z

Rodrigo Berriel: add autores

Ego-Lane Analysis System

2016-04-04T13:56:55Z

Rodrigo Berriel: adicionei link para vídeo no Youtube

[[category:Publicações]]
''This paper is under review (Special Issue: Automotive Vision, Image and Vision Computing)''

Adicionar Graphical Abstract

== Abstract ==

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines, Kalman filter and Particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e. lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.

== Demonstration Video ==
[https://www.youtube.com/watch?v=NPU9tiyA8vw Demonstration Video of ELAS]

== Dataset ==
Available upon acceptance.

== Source-Code ==
Available upon acceptance.

Ego-Lane Analysis System

2016-04-03T20:40:55Z

Rodrigo Berriel: Criação da página para gerar o link

[[category:Publicações]]
''This paper is under review (Special Issue: Automotive Vision, Image and Vision Computing)''

Adicionar Graphical Abstract

== Abstract ==

Decreasing costs of vision sensors and advances in embedded hardware boosted lane related research – detection, estimation, tracking, etc. – in the past two decades. The interest in this topic has increased even more with the demand for advanced driver assistance systems (ADAS) and self-driving cars. Although extensively studied independently, there is still need for studies that propose a combined solution for the multiple problems related to the ego-lane, such as lane departure warning (LDW), lane change detection, lane marking type (LMT) classification, road markings detection and classification, and detection of adjacent lanes presence. In this paper, we propose a real-time Ego-Lane Analysis System (ELAS) capable of estimating ego-lane position, classifying LMTs and road markings, performing LDW and detecting lane change events. The proposed vision-based system works on a temporal sequence of images. Lane marking features are extracted in perspective and Inverse Perspective Mapping (IPM) images that are combined to increase robustness. The final estimated lane is modeled as a spline using a combination of methods (Hough lines, Kalman filter and Particle filter). Based on the estimated lane, all other events are detected. To validate ELAS and cover the lack of lane datasets in the literature, a new dataset with more than 20 different scenes (in more than 15,000 frames) and considering a variety of scenarios (urban road, highways, traffic, shadows, etc.) was created. The dataset was manually annotated and made publicly available to enable evaluation of several events that are of interest for the research community (i.e. lane estimation, change, and centering; road markings; intersections; LMTs; crosswalks and adjacent lanes). Moreover, the system was also validated quantitatively and qualitatively on other public datasets. ELAS achieved high detection rates in all real-world events and proved to be ready for real-time applications.

== Video Demonstration ==
Adicionar Link para Vídeo

== Dataset ==
Available upon acceptance.

== Source-Code ==
Available upon acceptance.

A Particle Filter-based Lane Marker Tracking Approach using a Cubic Spline Model

2016-04-01T15:04:08Z

Rodrigo Berriel: adicionei página na categoria de publicações

[[category:Publicações]]
Conference Paper, SIBGRAPI, 2015

DOI: [http://dx.doi.org/10.1109/SIBGRAPI.2015.15 10.1109/SIBGRAPI.2015.15]

== Authors ==
Rodrigo Berriel, Edilson de Aguiar, Vanderlei Vieira de Souza Filho, Thiago Oliveira-Santos

== Abstract ==
In this paper we present a particle filter-based lane marker tracking approach using a cubic spline model. The system can detect the two main lane markers (i.e. lane strips) of marked roads from a monocular camera mounted on the top of a vehicle. Traditional lane marker detection and tracking systems have limitations to properly detect curved roads and to use temporal information to better estimate and track lane markers. The proposed system works on a temporal sequence of images. For each image, one at time, it applies a sequence of steps comprising an inverse perspective mapping to correct for perspective distortions, and a particle filter to smoothly track the lane markers along time. The output of the system is a lane marker generated by a cubic spline interpolation scheme to fit a wider range of lanes. Our system can run in real applications and it was validated with various road and traffic conditions. As a result, it achieves a high precision (98.13%) and a small error (0.0143 meters).

[[imagem:Resumo_A_Particle_Filter-based_Lane_Marker_Tracking_Approach_using_a_Cubic_Spline_Model.jpg]]

== Video ==
[https://youtu.be/cI79RZhIDCI Demo]

== BibTeX ==

@INPROCEEDINGS{Berriel2015,
author = {Rodrigo Ferreira Berriel and Edilson de Aguiar and Vanderlei Vieira de Souza Filho and Thiago Oliveira-Santos},
booktitle = {Graphics, Patterns and Images (SIBGRAPI), 2015 28th SIBGRAPI Conference on},
title = {A Particle Filter-Based Lane Marker Tracking Approach Using a Cubic Spline Model},
year = {2015},
month = {Aug},
pages = {149-156},
doi = {10.1109/SIBGRAPI.2015.15}
}

A Particle Filter-based Lane Marker Tracking Approach using a Cubic Spline Model

2016-04-01T14:46:42Z

Rodrigo Berriel: Adicionei o DOI e o BibTeX

Conference Paper, SIBGRAPI, 2015

DOI: [http://dx.doi.org/10.1109/SIBGRAPI.2015.15 10.1109/SIBGRAPI.2015.15]

== Authors ==
Rodrigo Berriel, Edilson de Aguiar, Vanderlei Vieira de Souza Filho, Thiago Oliveira-Santos

== Abstract ==
In this paper we present a particle filter-based lane marker tracking approach using a cubic spline model. The system can detect the two main lane markers (i.e. lane strips) of marked roads from a monocular camera mounted on the top of a vehicle. Traditional lane marker detection and tracking systems have limitations to properly detect curved roads and to use temporal information to better estimate and track lane markers. The proposed system works on a temporal sequence of images. For each image, one at time, it applies a sequence of steps comprising an inverse perspective mapping to correct for perspective distortions, and a particle filter to smoothly track the lane markers along time. The output of the system is a lane marker generated by a cubic spline interpolation scheme to fit a wider range of lanes. Our system can run in real applications and it was validated with various road and traffic conditions. As a result, it achieves a high precision (98.13%) and a small error (0.0143 meters).

[[imagem:Resumo_A_Particle_Filter-based_Lane_Marker_Tracking_Approach_using_a_Cubic_Spline_Model.jpg]]

== Video ==
[https://youtu.be/cI79RZhIDCI Demo]

== BibTeX ==

@INPROCEEDINGS{Berriel2015,
author = {Rodrigo Ferreira Berriel and Edilson de Aguiar and Vanderlei Vieira de Souza Filho and Thiago Oliveira-Santos},
booktitle = {Graphics, Patterns and Images (SIBGRAPI), 2015 28th SIBGRAPI Conference on},
title = {A Particle Filter-Based Lane Marker Tracking Approach Using a Cubic Spline Model},
year = {2015},
month = {Aug},
pages = {149-156},
doi = {10.1109/SIBGRAPI.2015.15}
}

Equipe

2016-03-31T22:06:10Z

Rodrigo Berriel: /* Add Rodrigo Berriel */

__TOC__

==Researchers==

* [http://www.lcad.inf.ufes.br/team/index.php/Dr._Alberto_Ferreira_De_Souza Dr. Alberto Ferreira De Souza (Coordinator)]
* Dr. Andrea Maria Pedrosa Valli
* Dr. Claudine Badue
* Dr. Edilson de Aguiar
* Dr. Elias Oliveira
* [http://www.lcad.inf.ufes.br/team/index.php/Dr._Fabio_Daros_Freitas Dr. Fabio Daros Freitas]
* Dr. Lucia Catabriga
* Dr. Maria Claudia Silva Boeres
* Dr. Maria Cristina Rangel
* Dr. Thiago Oliveira dos Santos

==Phd Students==

* [http://www.lcad.inf.ufes.br/~avelino Avelino Forechi]
* [http://www.lcad.inf.ufes.br/~lveronese Lucas de Paula Veronese]
* [http://www.inf.ufes.br/~mberger Mariella Berger]

==Masters Students==

* Cayo Fontana
* [http://www.lcad.inf.ufes.br/~filipe Filipe Wall Mutz]
* Lauro Jose Lyrio Junior
* Michael André Goncalves
* Romulo Ramos Radaelli
* [http://www.lcad.inf.ufes.br/~toliveira Tiago Alves de Oliveira]
* Vitor Barbirato
* [http://rodrigoberriel.com Rodrigo Berriel]

==Graduate Students==

* Lucas Catabriga
* Rafael Correia Nascimento
* Leornado Cunha
* Ranick Guidolini