What is it about?
A new paradigm of structured learning based vehicle behavior analysis has been proposed in this research project. Specifically, structured labels, which are capable of capturing the most prominent visual information of vehicle behaviors, are adopted instead of the traditional discrete numeric labels in predicting vehicle behaviors so that visually meaningful results can be outputted.
Featured Image
Photo by Kim Daniel on Unsplash
Why is it important?
Structured labels will be combined with deep forest for the first time, to learn the behavior of vehicles. Instead of computing the trajectory of a vehicle, a few images or even a single image is enough for determine the behavior of the vehicle.
Perspectives
The research results of this paper will enhance the ability of environment perception by automated vehicles and intelligent surveillance, and finally contribute to the transportation safety.
Luntian Mou
Peking University
Read the Original
This page is a summary of: Structured Behavior Prediction of On-road Vehicles via Deep Forest, Electronics Letters, March 2019, the Institution of Engineering and Technology (the IET),
DOI: 10.1049/el.2019.0472.
You can read the full text:
Contributors
The following have contributed to this page
