Method

Occlusion Patterns Deformable Part Model [OC-DPM]


Submitted on 16 Nov. 2013 19:57 by
Bojan Pepikj (Max Planck Institute for Informatics)

Running time:10 s
Environment:8 cores @ 2.5 Ghz (Matlab)

Method Description:
OC-DPM model with explicit occlusion handling.
Parameters:
C = 0.001, 12 viewpoint bins.
Latex Bibtex:
@inproceedings {Pepik2013CVPR,
title = {Occlusion Patterns for Object Class Detection},
booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
year = {2013},
publisher = {IEEE},
organization = {IEEE},
address = {Portland, OR, USA},
attachments = {http://www.d2.mpi-inf.mpg.de/sites/default/files/paper_final.pdf},
author = {Pepik, Bojan and Michael Stark and Peter Gehler and Schiele, Bernt}
}

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) 79.07 % 67.06 % 52.61 %
Car (Orientation) 77.35 % 65.32 % 51.00 %
This table as LaTeX


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



Orientation estimation results.
This figure as: png eps pdf txt gnuplot




eXTReMe Tracker