Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces RAPNet, a Graph Neural Network (GNN) designed to solve the ranked assignment problem within the δ-Generalized Labeled Multi-Bernoulli (δ-GLMB) filter for Multi-Object Tracking (MOT). The ranked assignment problem arises during track hypothesis truncation, a necessary step to manage computational complexity in MOT. Experimental results demonstrate that RAPNet achieves higher accuracy compared to the Gibbs sampling approach, offering a potential improvement over existing methods.
GNNs can outperform traditional methods like Gibbs sampling in the ranked assignment problem for multi-object tracking, offering a path to more accurate data association.
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 δ-Generalized Labeled Multi-Bernoulli (δ-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.