On the detection-to-track association for online multi-object tracking

TL;DR

This paper introduces a hybrid track association algorithm that leverages historical appearance distance data modeled by Gaussian mixtures to enhance detection-to-track association in online multi-object tracking, improving accuracy with minimal speed loss.

Contribution

It proposes a novel hybrid association method using incremental Gaussian mixture models to utilize historical appearance data, advancing multi-object tracking accuracy.

Findings

HTA improves target identification performance.

DeepSORT with HTA achieves better or comparable results.

Small speed compromise for significant accuracy gains.

Abstract

Driven by recent advances in object detection with deep neural networks, the tracking-by-detection paradigm has gained increasing prevalence in the research community of multi-object tracking (MOT). It has long been known that appearance information plays an essential role in the detection-to-track association, which lies at the core of the tracking-by-detection paradigm. While most existing works consider the appearance distances between the detections and the tracks, they ignore the statistical information implied by the historical appearance distance records in the tracks, which can be particularly useful when a detection has similar distances with two or more tracks. In this work, we propose a hybrid track association (HTA) algorithm that models the historical appearance distances of a track with an incremental Gaussian mixture model (IGMM) and incorporates the derived statistical information into the calculation of the detection-to-track association cost. Experimental results on three MOT benchmarks confirm that HTA effectively improves the target identification performance with a small compromise to the tracking speed. Additionally, compared to many state-of-the-art trackers, the DeepSORT tracker equipped with HTA achieves better or comparable performance in terms of the balance of tracking quality and speed.