TAPTRv2: Attention-based Position Update Improves Tracking Any Point

TL;DR

TAPTRv2 introduces an attention-based position update mechanism that enhances transformer-based tracking of any point, eliminating reliance on cost-volume and achieving state-of-the-art results.

Contribution

It proposes a novel attention-based position update operation using key-aware deformable attention, improving tracking accuracy and efficiency over previous TAPTR models.

Findings

Surpasses TAPTR in performance on multiple datasets

Removes the need for cost-volume computation

Achieves state-of-the-art tracking accuracy

Abstract

In this paper, we present TAPTRv2, a Transformer-based approach built upon TAPTR for solving the Tracking Any Point (TAP) task. TAPTR borrows designs from DEtection TRansformer (DETR) and formulates each tracking point as a point query, making it possible to leverage well-studied operations in DETR-like algorithms. TAPTRv2 improves TAPTR by addressing a critical issue regarding its reliance on cost-volume,which contaminates the point query\'s content feature and negatively impacts both visibility prediction and cost-volume computation. In TAPTRv2, we propose a novel attention-based position update (APU) operation and use key-aware deformable attention to realize. For each query, this operation uses key-aware attention weights to combine their corresponding deformable sampling positions to predict a new query position. This design is based on the observation that local attention is essentially the same as cost-volume, both of which are computed by dot-production between a query and its surrounding features. By introducing this new operation, TAPTRv2 not only removes the extra burden of cost-volume computation, but also leads to a substantial performance improvement. TAPTRv2 surpasses TAPTR and achieves state-of-the-art performance on many challenging datasets, demonstrating the superiority