Robust Multi-Object Tracking by Marginal Inference

TL;DR

This paper introduces a real-time method to compute stable marginal probabilities for object pairs in multi-object tracking, enabling a single threshold approach that improves IDF1 scores and interpretability across different videos.

Contribution

The paper proposes a novel marginal inference method that stabilizes distance measures, allowing consistent thresholding and enhancing multi-object tracking performance.

Findings

Achieves about 1 point improvement in IDF1 metric.

Provides competitive results on MOT17 and MOT20 benchmarks.

Offers more interpretable probabilities for post-processing.

Abstract

Multi-object tracking in videos requires to solve a fundamental problem of one-to-one assignment between objects in adjacent frames. Most methods address the problem by first discarding impossible pairs whose feature distances are larger than a threshold, followed by linking objects using Hungarian algorithm to minimize the overall distance. However, we find that the distribution of the distances computed from Re-ID features may vary significantly for different videos. So there isn't a single optimal threshold which allows us to safely discard impossible pairs. To address the problem, we present an efficient approach to compute a marginal probability for each pair of objects in real time. The marginal probability can be regarded as a normalized distance which is significantly more stable than the original feature distance. As a result, we can use a single threshold for all videos. The approach is general and can be applied to the existing trackers to obtain about one point improvement in terms of IDF1 metric. It achieves competitive results on MOT17 and MOT20 benchmarks. In addition, the computed probability is more interpretable which facilitates subsequent post-processing operations.