ClickTrack: Towards Real-time Interactive Single Object Tracking

TL;DR

ClickTrack introduces a real-time interactive single object tracking method that uses clicking and textual guidance to improve initialization and reduce ambiguity, enhancing performance on benchmarks and integrating with SAM.

Contribution

The paper proposes ClickTrack, a novel paradigm using clicking interaction and guided refinement for improved real-time single object tracking.

Findings

Achieves stable real-time tracking performance on LaSOT and GOT-10k.

Reduces ambiguity in SAM when using point inputs.

Demonstrates effective integration of GCR with existing models.

Abstract

Single object tracking(SOT) relies on precise object bounding box initialization. In this paper, we reconsidered the deficiencies in the current approaches to initializing single object trackers and propose a new paradigm for single object tracking algorithms, ClickTrack, a new paradigm using clicking interaction for real-time scenarios. Moreover, click as an input type inherently lack hierarchical information. To address ambiguity in certain special scenarios, we designed the Guided Click Refiner(GCR), which accepts point and optional textual information as inputs, transforming the point into the bounding box expected by the operator. The bounding box will be used as input of single object trackers. Experiments on LaSOT and GOT-10k benchmarks show that tracker combined with GCR achieves stable performance in real-time interactive scenarios. Furthermore, we explored the integration of GCR into the Segment Anything model(SAM), significantly reducing ambiguity issues when SAM receives point inputs.