Enhancing real-time human detection based on histograms of oriented gradients

In this paper we propose a human detection framework based on an enhanced version of Histogram of Oriented Gradients (HOG) features. These feature descriptors are computed with the help of a precalculated histogram of square-blocks. This novel method outperforms the integral of oriented histograms a...

ver descrição completa

Detalhes bibliográficos
Autores: Perdersoli, Marco, Gonzàlez, Jordi, Chakraborty, Bhaskar, Villanueva, Juan J.
Formato: capítulo de livro
Fecha de publicación:2007
País:España
Recursos:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/2675
Acesso em linha:https://hdl.handle.net/2117/2675
Access Level:acceso abierto
Palavra-chave:Computer vision
Visió per ordinador
Classificació INSPEC::Pattern recognition::Computer vision
Àrees temàtiques de la UPC::Enginyeria de la telecomunicació::Processament del senyal::Processament de la imatge i del senyal vídeo
Descrição
Resumo:In this paper we propose a human detection framework based on an enhanced version of Histogram of Oriented Gradients (HOG) features. These feature descriptors are computed with the help of a precalculated histogram of square-blocks. This novel method outperforms the integral of oriented histograms allowing the calculation of a single feature four times faster. Using Adaboost for HOG feature selection and Support Vector Machine as weak classifier, we build up a real-time human classifier with an excellent detection rate.