AttTrack: Online Deep Attention Transfer for Multi-object Tracking

TL;DR

AttTrack introduces an online deep attention transfer method that enhances lightweight multi-object tracking models by leveraging high-level features from complex networks, improving accuracy with minimal speed loss.

Contribution

The paper proposes a novel online knowledge transfer framework for multi-object tracking that combines feature alignment, model interleaving, and teacher predictions to boost lightweight model performance.

Findings

Significant accuracy improvements on MOT17 and MOT15 datasets.

Minor degradation in tracking speed with enhanced performance.

Effective knowledge transfer from complex to lightweight models.

Abstract

Multi-object tracking (MOT) is a vital component of intelligent video analytics applications such as surveillance and autonomous driving. The time and storage complexity required to execute deep learning models for visual object tracking hinder their adoption on embedded devices with limited computing power. In this paper, we aim to accelerate MOT by transferring the knowledge from high-level features of a complex network (teacher) to a lightweight network (student) at both training and inference times. The proposed AttTrack framework has three key components: 1) cross-model feature learning to align intermediate representations from the teacher and student models, 2) interleaving the execution of the two models at inference time, and 3) incorporating the updated predictions from the teacher model as prior knowledge to assist the student model. Experiments on pedestrian tracking tasks are conducted on the MOT17 and MOT15 datasets using two different object detection backbones YOLOv5 and DLA34 show that AttTrack can significantly improve student model tracking performance while sacrificing only minor degradation of tracking speed.