LoopTrack: A Simple Baseline for Parameter-Efficient Transformer Tracking

2026-09-27

AI Summary

Existing Transformer-based trackers suffer from parameter redundancy, hindering their deployment on resource-constrained devices. To address this, this work proposes LoopTrack, a framework that leverages a parameter-shared recurrent architecture to enable efficient feature interaction while significantly reducing model complexity. Methodologically, it introduces an object-aware recurrence with a gated object memory mechanism, integrated with lightweight attention modules to achieve progressive feature refinement and effectively mitigate temporal drift. Experimental results demonstrate that with only 3.4M parameters, the proposed model attains a success score of 66.2% on the LaSOT benchmark, outperforming existing methods of comparable scale. Consequently, this study establishes a strong baseline for lightweight visual tracking.

Abstract

Current Transformer-based tracking methods typically stack multiple Transformer blocks with separate parameters to model interactions between the target template and the search region for target localization. These trackers often incur substantial parameter overhead from stacked blocks, making their deployment on resource-limited devices difficult. To address this, we propose a parameter-efficient Transformer tracking framework, dubbed LoopTrack, which repeatedly applies a set of Transformer blocks with shared parameters to interact features in a looped architecture for tracking, significantly reducing the number of parameters. To further exploit target cues, we present two lightweight designs, including target-aware looping (TAL) and gated target memory (GTM). The former applies intermediate target information generated by one loop to guide feature interaction in the subsequent loop, enabling progressive feature refinement, while the latter maintains a compact memory across frames, which is incorporated into the loop process to provide long-term information to the tracker, mitigating temporal drift in tracking. Compared to existing Transformer trackers, LoopTrack enables multiple rounds of feature interaction with fewer model parameters, making it resource-friendly for deployment. In extensive experiments on multiple datasets, LoopTrack shows a favorable accuracy-parameter trade-off. In particular, our LoopTrack$_{\rm One}$, with a single shared Transformer block, achieves 66.2\% SUC score on LaSOT with only 3.4M parameters, while LoopTrack$_{\rm Three}$, using three shared blocks, achieves 69.3\% SUC score with 6.4M parameters, surpassing existing parameter-efficient tracking methods with comparable or larger model size. With LoopTrack, we aim to establish a simple yet strong baseline for parameter-efficient Transformer tracking. Our code and models will be released.

Problem

• Visual Tracking
• Parameter Efficiency
• Transformer
• Resource-constrained Deployment

Innovation

• Parameter-Efficient Tracking
• Shared Parameters
• Looped Architecture
• Target-Aware Looping
• Gated Target Memory