Method

Reinspect [Reinspect]
http://github.com/russell91/tensorbox

Submitted on 22 May. 2016 22:31 by
russell stewart (Stanford)

Running time:2s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
OverFeat using GoogLeNet and LSTM decoder for
occlusions. Attention with rezooming is added to the original
publication.
Parameters:
\alpha = 0.2
Latex Bibtex:
@inproceedings{Stewart2016CVPR,
author = {Russell Stewart and Mykhaylo Andriluka and Andrew Y. Ng},
title = {End-to-End People Detection in Crowded Scenes},
booktitle = CVPR,
year = {2016}
}

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
Car (Detection) 90.27 % 77.48 % 66.73 %
This table as LaTeX

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




eXTReMe Tracker