Method

Deep Learning Strong Parts for Pedestrian Detection [DeepParts]


Submitted on 30 Jul. 2015 19:20 by
Yonglong Tian (The Chinese University of Hong Kong)

Running time:~1 s
Environment:GPU @ 2.5 Ghz (Matlab)

Method Description:
Parameters:
Latex Bibtex:
@inproceedings{Tian2015ICCV,
title = {Deep Learning Strong Parts for Pedestrian
Detection},
author = {Yonglong Tian and Ping Luo and Xiaogang
Wang and Xiaoou Tang},
booktitle = {ICCV},
year = {2015}
}

Detailed Results

Object detection and orientation estimation results. Results for object detection are given in terms of average precision (AP) and results for joint object detection and orientation estimation are provided in terms of average orientation similarity (AOS).


Benchmark Easy Moderate Hard
Pedestrian (Detection) 71.47 % 58.15 % 51.92 %
This table as LaTeX


2D object detection results.
This figure as: png eps pdf txt gnuplot




eXTReMe Tracker