PointTrackNet: An End-to-End Network For 3-D Object Detection and Tracking From Point Clouds

TL;DR

PointTrackNet is an end-to-end 3D object detection and tracking network that uses only two adjacent point-cloud frames to improve tracking accuracy in rapid and irregular scenarios.

Contribution

It introduces a novel end-to-end network that jointly performs detection and tracking from point clouds, overcoming limitations of traditional filter-based methods.

Findings

Achieves competitive results on KITTI dataset.

Performs well in irregular and rapid motion scenarios.

Outperforms some existing methods in challenging conditions.

Abstract

Recent machine learning-based multi-object tracking (MOT) frameworks are becoming popular for 3-D point clouds. Most traditional tracking approaches use filters (e.g., Kalman filter or particle filter) to predict object locations in a time sequence, however, they are vulnerable to extreme motion conditions, such as sudden braking and turning. In this letter, we propose PointTrackNet, an end-to-end 3-D object detection and tracking network, to generate foreground masks, 3-D bounding boxes, and point-wise tracking association displacements for each detected object. The network merely takes as input two adjacent point-cloud frames. Experimental results on the KITTI tracking dataset show competitive results over the state-of-the-arts, especially in the irregularly and rapidly changing scenarios.