Method

3D Voxel Patterns [3DVP]
https://github.com/yuxng/3DVP

Submitted on 13 Nov. 2014 09:15 by
Yu Xiang (Stanford University)

Running time:40 s
Environment:8 cores @ 3.5 Ghz (Matlab + C/C++)

Method Description:
Data-Driven 3D Voxel Patterns for Object Category
Recognition

http://cvgl.stanford.edu/projects/3DVP/
Parameters:
Please see our code for details.
Latex Bibtex:
@InProceedings{Xiang2015CVPR,
author = {Yu Xiang and Wongun Choi and
Yuanqing Lin and Silvio Savarese},
title = {Data-Driven 3D Voxel Patterns
for Object Category Recognition},
booktitle = {IEEE Conference on Computer
Vision and Pattern Recognition},
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
Car (Detection) 84.95 % 76.98 % 65.78 %
Car (Orientation) 84.44 % 75.71 % 64.41 %
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