Method

Extension and Attention Network [ExtAtt]


Submitted on 19 Jan. 2019 13:32 by
zhihang Fu (ZJU)

Running time:1.2 s
Environment:GPU @ 2.5 Ghz (Python + C/C++)

Method Description:
The first stage produces proposals by multiple output
layers, each focusing on instances within certain scale
ranges; next the extension and attention stage refines
the results with multiple loss functions.
Parameters:
TBD
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
Pedestrian (Detection) 87.95 % 79.63 % 74.78 %
Cyclist (Detection) 84.04 % 74.25 % 65.03 %
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




eXTReMe Tracker