Method

SINet: A Scale-Insensitive Convolutional Neural Network for Fast Vehicle Detection [SINet_VGG]


Submitted on 10 Jun. 2017 10:17 by
Xiaowei HU (The Chinese University of Hong Kong)

Running time:0.2 s
Environment:GPU @ 2.5 Ghz (Matlab + C/C++)

Method Description:
Paper and code will be released later.
GPU: TITAN X (Not Pascal version)
Parameters:
Network: VGG
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) 90.60 % 89.56 % 78.19 %
This table as LaTeX


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




eXTReMe Tracker