Method

You only lool once [YOLO]


Submitted on 31 Mar. 2017 01:26 by
Guangrui Liu (Texas A&M University - Corpus Christi)

Running time:0.03 s
Environment:GPU @ 1.0 Ghz (C/C++)

Method Description:
Train YOLO on KITTI dataset
Parameters:
thresh=0.2
Latex Bibtex:

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) 49.47 % 35.86 % 29.74 %
Pedestrian (Detection) 25.63 % 24.35 % 17.50 %
Cyclist (Detection) 18.07 % 13.96 % 13.83 %
This table as LaTeX


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



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



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




eXTReMe Tracker