Deep HM-SORT: Enhancing Multi-Object Tracking in Sports with Deep Features, Harmonic Mean, and Expansion IOU

Matias Gran-Henriksen,Hans Andreas Lindgaard,Gabriel Kiss,Frank Lindseth

2024-06-18

Abstract:This paper introduces Deep HM-SORT, a novel online multi-object tracking algorithm specifically designed to enhance the tracking of athletes in sports scenarios. Traditional multi-object tracking methods often struggle with sports environments due to the similar appearances of players, irregular and unpredictable movements, and significant camera motion. Deep HM-SORT addresses these challenges by integrating deep features, harmonic mean, and Expansion IOU. By leveraging the harmonic mean, our method effectively balances appearance and motion cues, significantly reducing ID-swaps. Additionally, our approach retains all tracklets indefinitely, improving the re-identification of players who leave and re-enter the frame. Experimental results demonstrate that Deep HM-SORT achieves state-of-the-art performance on two large-scale public benchmarks, SportsMOT and SoccerNet Tracking Challenge 2023. Specifically, our method achieves 80.1 HOTA on the SportsMOT dataset and 85.4 HOTA on the SoccerNet-Tracking dataset, outperforming existing trackers in key metrics such as HOTA, IDF1, AssA, and MOTA. This robust solution provides enhanced accuracy and reliability for automated sports analytics, offering significant improvements over previous methods without introducing additional computational cost.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper aims to address the problem of multi-object tracking in sports scenarios. Traditional multi-object tracking methods perform poorly in sports environments due to factors such as the similar appearance of athletes, irregular and unpredictable movements, and significant camera motion. Specifically, the paper proposes the Deep HM-SORT algorithm, which integrates deep features, harmonic mean, and Expansion IOU to solve the following key issues:

1. **Reducing ID switches**: By combining appearance and motion cues, it effectively balances the relationship between the two, significantly reducing ID switches.
2. **Enhancing re-identification capability**: It retains all trajectory fragments, ensuring that athletes can be correctly re-identified even if they temporarily leave the frame and return.
3. **Improving tracking accuracy and reliability**: It achieves state-of-the-art performance on two large-scale public benchmark datasets (SportsMOT and SoccerNet Tracking Challenge 2023), particularly outperforming existing methods in key metrics such as HOTA, IDF1, AssA, and MOTA.

Overall, Deep HM-SORT aims to provide a more accurate and reliable solution for automated sports analysis, especially optimized for the complex movements of athletes on the field.