Leveraging Foundation Models via Knowledge Distillation in Multi-Object Tracking: Distilling DINOv2 Features to FairMOT

TL;DR

This paper explores using a large foundation model, DINOv2, to enhance multi-object tracking by distilling its features into a traditional model, aiming to improve performance without extensive retraining.

Contribution

It introduces a novel knowledge distillation approach from DINOv2 to FairMOT, demonstrating how foundation models can be integrated into existing tracking frameworks.

Findings

Partial performance improvements in specific scenarios

Limitations in consistently outperforming baseline models

Insights into foundation models' applicability in MOT

Abstract

Multiple Object Tracking (MOT) is a computer vision task that has been employed in a variety of sectors. Some common limitations in MOT are varying object appearances, occlusions, or crowded scenes. To address these challenges, machine learning methods have been extensively deployed, leveraging large datasets, sophisticated models, and substantial computational resources. Due to practical limitations, access to the above is not always an option. However, with the recent release of foundation models by prominent AI companies, pretrained models have been trained on vast datasets and resources using state-of-the-art methods. This work tries to leverage one such foundation model, called DINOv2, through using knowledge distillation. The proposed method uses a teacher-student architecture, where DINOv2 is the teacher and the FairMOT backbone HRNetv2 W18 is the student. The results imply that although the proposed method shows improvements in certain scenarios, it does not consistently outperform the original FairMOT model. These findings highlight the potential and limitations of applying foundation models in knowledge