CLDTracker: A Comprehensive Language Description for Visual Tracking

TL;DR

CLDTracker introduces a dual-branch framework leveraging rich, temporally-adaptive language descriptions and visual features, achieving state-of-the-art results in visual object tracking by effectively integrating semantic and visual cues.

Contribution

The paper presents a novel dual-branch architecture that constructs comprehensive textual descriptions using VLMs, enhancing visual tracking performance.

Findings

Achieves SOTA performance on six VOT benchmarks.

Effectively integrates semantic and visual features for robust tracking.

Demonstrates the benefits of temporally-adaptive language representations.

Abstract

VOT remains a fundamental yet challenging task in computer vision due to dynamic appearance changes, occlusions, and background clutter. Traditional trackers, relying primarily on visual cues, often struggle in such complex scenarios. Recent advancements in VLMs have shown promise in semantic understanding for tasks like open-vocabulary detection and image captioning, suggesting their potential for VOT. However, the direct application of VLMs to VOT is hindered by critical limitations: the absence of a rich and comprehensive textual representation that semantically captures the target object's nuances, limiting the effective use of language information; inefficient fusion mechanisms that fail to optimally integrate visual and textual features, preventing a holistic understanding of the target; and a lack of temporal modeling of the target's evolving appearance in the language domain, leading to a disconnect between the initial description and the object's subsequent visual changes. To bridge these gaps and unlock the full potential of VLMs for VOT, we propose CLDTracker, a novel Comprehensive Language Description framework for robust visual Tracking. Our tracker introduces a dual-branch architecture consisting of a textual and a visual branch. In the textual branch, we construct a rich bag of textual descriptions derived by harnessing the powerful VLMs such as CLIP and GPT-4V, enriched with semantic and contextual cues to address the lack of rich textual representation. Experiments on six standard VOT benchmarks demonstrate that CLDTracker achieves SOTA performance, validating the effectiveness of leveraging robust and temporally-adaptive vision-language representations for tracking. Code and models are publicly available at: https://github.com/HamadYA/CLDTracker