MV-TAP: Tracking Any Point in Multi-View Videos

TL;DR

MV-TAP is a new multi-view point tracking method that uses camera geometry and cross-view attention to improve trajectory estimation in complex scenes, supported by synthetic and real-world datasets.

Contribution

Introduces MV-TAP, a novel multi-view point tracker leveraging cross-view attention and geometry, with a large synthetic dataset for training and evaluation.

Findings

MV-TAP outperforms existing methods on benchmarks.

The approach improves trajectory completeness and reliability.

Extensive experiments validate its effectiveness.

Abstract

Multi-view camera systems enable rich observations of complex real-world scenes, and understanding dynamic objects in multi-view settings has become central to various applications. In this work, we present MV-TAP, a novel point tracker that tracks points across multi-view videos of dynamic scenes by leveraging cross-view information. MV-TAP utilizes camera geometry and a cross-view attention mechanism to aggregate spatio-temporal information across views, enabling more complete and reliable trajectory estimation in multi-view videos. To support this task, we construct a large-scale synthetic training dataset and real-world evaluation sets tailored for multi-view tracking. Extensive experiments demonstrate that MV-TAP outperforms existing point-tracking methods on challenging benchmarks, establishing an effective baseline for advancing research in multi-view point tracking.