Tracking Anything with Decoupled Video Segmentation

TL;DR

This paper introduces DEVA, a decoupled video segmentation method that combines task-specific image segmentation with universal temporal propagation, enabling flexible and cost-effective tracking of various objects without extensive task-specific training.

Contribution

The paper proposes a novel decoupled approach for video segmentation that separates image-level segmentation from temporal propagation, reducing training costs and improving generalization across tasks.

Findings

Outperforms end-to-end methods in data-scarce scenarios

Effective in large-vocabulary and open-world segmentation tasks

Enables task-specific segmentation with a universal temporal model

Abstract

Training data for video segmentation are expensive to annotate. This impedes extensions of end-to-end algorithms to new video segmentation tasks, especially in large-vocabulary settings. To 'track anything' without training on video data for every individual task, we develop a decoupled video segmentation approach (DEVA), composed of task-specific image-level segmentation and class/task-agnostic bi-directional temporal propagation. Due to this design, we only need an image-level model for the target task (which is cheaper to train) and a universal temporal propagation model which is trained once and generalizes across tasks. To effectively combine these two modules, we use bi-directional propagation for (semi-)online fusion of segmentation hypotheses from different frames to generate a coherent segmentation. We show that this decoupled formulation compares favorably to end-to-end approaches in several data-scarce tasks including large-vocabulary video panoptic segmentation, open-world video segmentation, referring video segmentation, and unsupervised video object segmentation. Code is available at: https://hkchengrex.github.io/Tracking-Anything-with-DEVA