MambaMOT: State-Space Model as Motion Predictor for Multi-Object Tracking

TL;DR

This paper introduces MambaMOT, a learning-based motion prediction model that surpasses traditional Kalman filter methods in multi-object tracking, especially in complex, nonlinear, and occlusion-heavy scenarios like sports and dance.

Contribution

It proposes replacing the Kalman filter with a neural network-based motion model, improving tracking accuracy and robustness in challenging environments.

Findings

Outperforms traditional methods on DanceTrack and SportsMOT datasets.

Handles complex, nonlinear motions more effectively.

Improves tracking robustness during occlusions.

Abstract

In the field of multi-object tracking (MOT), traditional methods often rely on the Kalman filter for motion prediction, leveraging its strengths in linear motion scenarios. However, the inherent limitations of these methods become evident when confronted with complex, nonlinear motions and occlusions prevalent in dynamic environments like sports and dance. This paper explores the possibilities of replacing the Kalman filter with a learning-based motion model that effectively enhances tracking accuracy and adaptability beyond the constraints of Kalman filter-based tracker. In this paper, our proposed method MambaMOT and MambaMOT+, demonstrate advanced performance on challenging MOT datasets such as DanceTrack and SportsMOT, showcasing their ability to handle intricate, non-linear motion patterns and frequent occlusions more effectively than traditional methods.