Histograms of Oriented Gradients for Human Detection
N. Dalal, and B. Triggs. Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer
Society Conference on, (2005)
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.
%0 Journal Article
%1 citeulike:335784
%A Dalal, N.
%A Triggs, B.
%D 2005
%J Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer
Society Conference on
%K detection, vision
%P 886--893
%T Histograms of Oriented Gradients for Human Detection
%U http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1467360
%V 1
%X 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.
@article{citeulike:335784,
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.},
added-at = {2009-08-24T22:23:57.000+0200},
author = {Dalal, N. and Triggs, B.},
biburl = {https://www.bibsonomy.org/bibtex/23765f5c9ab9abdf0efe63d50c64147d5/sitrke},
citeulike-article-id = {335784},
citeulike-linkout-0 = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1467360},
description = {Human Detection},
file = {:paperpool\\hog.pdf:PDF},
interhash = {a19602b4d8fd42e73d957ae456280c17},
intrahash = {3765f5c9ab9abdf0efe63d50c64147d5},
journal = {Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer
Society Conference on},
keywords = {detection, vision},
pages = {886--893},
posted-at = {2007-03-13 17:24:40},
priority = {2},
timestamp = {2009-08-24T22:23:57.000+0200},
title = {Histograms of Oriented Gradients for Human Detection},
url = {http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1467360},
volume = 1,
year = 2005
}