Method

Weak Segmentation Supervised Deep Neural Networks [la] [WSSN]


Submitted on 31 Aug. 2020 15:33 by
Zhixin Guo (Ghent University)

Running time:0.37 s
Environment:GPU @ >3.5 Ghz (Python + C/C++)

Method Description:
Jointly optimize segmentaion and detection tasks.
With RGBD input. Note that the generation time of
the depth map is not included in the overall
runtime.
Parameters:
see Paper.
Latex Bibtex:
@article{guo2021weak,
title={Weak Segmentation Supervised Deep Neural
Networks for Pedestrian Detection},
author={Guo, Zhixin and Liao, Wenzhi and Xiao,
Yifan and Veelaert, Peter and Philips, Wilfried},
journal={Pattern Recognition},
pages={108063},
year={2021},
publisher={Elsevier}
}

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) 84.91 % 76.42 % 71.86 %
This table as LaTeX


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




eXTReMe Tracker