MFTIQ: Multi-Flow Tracker with Independent Matching Quality Estimation

TL;DR

MFTIQ introduces an independent quality estimation module to enhance dense long-term point tracking in videos, improving accuracy and speed by decoupling correspondence quality from optical flow computations.

Contribution

It presents a plug-and-play framework that separates quality estimation from flow computation, enabling better accuracy and flexibility in long-term tracking.

Findings

Outperforms original MFT in accuracy on TAP-Vid Davis dataset.

Achieves comparable results to state-of-the-art trackers.

Offers faster processing speed with RoMa optical flow.

Abstract

In this work, we present MFTIQ, a novel dense long-term tracking model that advances the Multi-Flow Tracker (MFT) framework to address challenges in point-level visual tracking in video sequences. MFTIQ builds upon the flow-chaining concepts of MFT, integrating an Independent Quality (IQ) module that separates correspondence quality estimation from optical flow computations. This decoupling significantly enhances the accuracy and flexibility of the tracking process, allowing MFTIQ to maintain reliable trajectory predictions even in scenarios of prolonged occlusions and complex dynamics. Designed to be "plug-and-play", MFTIQ can be employed with any off-the-shelf optical flow method without the need for fine-tuning or architectural modifications. Experimental validations on the TAP-Vid Davis dataset show that MFTIQ with RoMa optical flow not only surpasses MFT but also performs comparably to state-of-the-art trackers while having substantially faster processing speed. Code and models available at https://github.com/serycjon/MFTIQ .