1st Place Solution for PVUW Challenge 2023: Video Panoptic Segmentation

TL;DR

This paper presents the winning solution for the PVUW 2023 challenge in video panoptic segmentation, utilizing a decoupling strategy to effectively leverage temporal information, achieving top performance scores.

Contribution

The paper introduces a decoupling strategy for video panoptic segmentation that improves the utilization of temporal information, validated through competitive results.

Findings

Achieved VPQ scores of 51.4 and 53.7 in development and test phases.

Ranked 1st in the VPS track of PVUW Challenge 2023.

Validated the effectiveness of the decoupling strategy.

Abstract

Video panoptic segmentation is a challenging task that serves as the cornerstone of numerous downstream applications, including video editing and autonomous driving. We believe that the decoupling strategy proposed by DVIS enables more effective utilization of temporal information for both "thing" and "stuff" objects. In this report, we successfully validated the effectiveness of the decoupling strategy in video panoptic segmentation. Finally, our method achieved a VPQ score of 51.4 and 53.7 in the development and test phases, respectively, and ultimately ranked 1st in the VPS track of the 2nd PVUW Challenge. The code is available at https://github.com/zhang-tao-whu/DVIS