Method

Lidar_ROI+Yolo(UJS) [Lidar_ROI+Yolo(UJS)]


Submitted on 8 Sep. 2018 09:31 by
Hai Wang (Jiangsu University)

Running time:0.1 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
The main framework is first use Lidar information
to generate ROI in image, and then use a improved
Yolo to detect and classifier object in image,
finally the detected object will be mapped to
Lidar coordinate again to verify if the detection
results are right.
Parameters:
The main framework is first use Lidar information
to generate ROI in image, and then use a improved
Yolo to detect and classifier object in image,
finally the detected object will be mapped to
Lidar coordinate again to verify if the detection
results are right.
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) 70.58 % 62.71 % 55.17 %
Car (Orientation) 28.93 % 25.40 % 22.51 %
Pedestrian (Detection) 47.11 % 38.76 % 32.33 %
Pedestrian (Orientation) 28.50 % 23.43 % 19.87 %
Cyclist (Detection) 39.41 % 27.21 % 26.12 %
Cyclist (Orientation) 13.88 % 9.31 % 9.12 %
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



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



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



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