A Graph Neural Network Approach for Solving the Ranked Assignment Problem in Multi-Object Tracking

TL;DR

This paper introduces RAPNet, a graph neural network model that improves the accuracy of solving the ranked assignment problem in multi-object tracking, surpassing traditional algorithms like Murty's and Gibbs sampling.

Contribution

The paper presents a novel GNN-based approach, RAPNet, for data association in MOT that enhances accuracy over existing methods.

Findings

RAPNet outperforms Gibbs sampling in accuracy.

RAPNet achieves comparable or better results than Murty's algorithm.

The approach leverages deep learning to model assignment problems effectively.

Abstract

Associating measurements with tracks is a crucial step in Multi-Object Tracking (MOT) to guarantee the safety of autonomous vehicles. To manage the exponentially growing number of track hypotheses, truncation becomes necessary. In the $\delta$-Generalized Labeled Multi-Bernoulli ($\delta$-GLMB) filter application, this truncation typically involves the ranked assignment problem, solved by Murty's algorithm or the Gibbs sampling approach, both with limitations in terms of complexity or accuracy, respectively. With the motivation to improve these limitations, this paper addresses the ranked assignment problem arising from data association tasks with an approach that employs Graph Neural Networks (GNNs). The proposed Ranked Assignment Prediction Graph Neural Network (RAPNet) uses bipartite graphs to model the problem, harnessing the computational capabilities of deep learning. The conclusive evaluation compares the RAPNet with Murty's algorithm and the Gibbs sampler, showing accuracy improvements compared to the Gibbs sampler.