3D Object Detection and Tracking Based on Streaming Data

TL;DR

This paper introduces a novel 3D object detection and tracking framework that leverages temporal information from streaming data, significantly improving detection accuracy and achieving competitive tracking results on benchmark datasets.

Contribution

It proposes a dual-way network for keyframe-based detection and motion-guided interpolation for non-key frames, enhancing 3D detection and tracking over traditional frame-by-frame methods.

Findings

Significant improvement in object detection accuracy over frame-by-frame methods

Achieved 76.68% MOTA and 81.65% MOTP on KITTI benchmark

Effective utilization of temporal information enhances 3D tracking performance

Abstract

Recent approaches for 3D object detection have made tremendous progresses due to the development of deep learning. However, previous researches are mostly based on individual frames, leading to limited exploitation of information between frames. In this paper, we attempt to leverage the temporal information in streaming data and explore 3D streaming based object detection as well as tracking. Toward this goal, we set up a dual-way network for 3D object detection based on keyframes, and then propagate predictions to non-key frames through a motion based interpolation algorithm guided by temporal information. Our framework is not only shown to have significant improvements on object detection compared with frame-by-frame paradigm, but also proven to produce competitive results on KITTI Object Tracking Benchmark, with 76.68% in MOTA and 81.65% in MOTP respectively.