Histograms of Oriented Gradients for Human Detection - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2005

Histograms of Oriented Gradients for Human Detection

Abstract

We study the question of feature sets for robust visual object recognition, adopting linear SVM based human detection as a test case. After reviewing existing edge and gradient based descriptors, we show experimentally that grids of Histograms of Oriented Gradient (HOG) descriptors significantly outperform existing feature sets for human detection. We study the influence of each stage of the computation on performance, concluding that fine-scale gradients, fine orientation binning, relatively coarse spatial binning, and high-quality local contrast normalization in overlapping descriptor blocks are all important for good results. The new approach gives near-perfect separation on the original MIT pedestrian database, so we introduce a more challenging dataset containing over 1800 annotated human images with a large range of pose variations and backgrounds.
Fichier principal
Vignette du fichier
hog_cvpr2005.pdf (445.45 Ko) Télécharger le fichier
cvpr2005_talk.pdf (567.43 Ko) Télécharger le fichier
demo.mpeg (41.71 Mo) Télécharger le fichier
Origin : Files produced by the author(s)
Format : Other
Format : Other

Dates and versions

inria-00548512 , version 1 (20-12-2010)

Identifiers

Cite

Navneet Dalal, Bill Triggs. Histograms of Oriented Gradients for Human Detection. International Conference on Computer Vision & Pattern Recognition (CVPR '05), Jun 2005, San Diego, United States. pp.886--893, ⟨10.1109/CVPR.2005.177⟩. ⟨inria-00548512⟩
21616 View
70238 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More