3rd Place Solution for PVUW Challenge 2024: Video Panoptic Segmentation

TL;DR

This paper presents a top-3 solution for video panoptic segmentation in the VIPSeg challenge, enhancing a baseline model with query-wise ensemble and techniques to improve VPQ scores.

Contribution

The paper introduces a novel approach combining query-wise ensemble with additional techniques to improve video panoptic segmentation performance.

Findings

Achieved a VPQ score of 57.01 on VIPSeg test set

Ranked 3rd in the VPS track of the challenge

Enhanced baseline model with ensemble and techniques

Abstract

Video panoptic segmentation is an advanced task that extends panoptic segmentation by applying its concept to video sequences. In the hope of addressing the challenge of video panoptic segmentation in diverse conditions, We utilize DVIS++ as our baseline model and enhance it by introducing a comprehensive approach centered on the query-wise ensemble, supplemented by additional techniques. Our proposed approach achieved a VPQ score of 57.01 on the VIPSeg test set, and ranked 3rd in the VPS track of the 3rd Pixel-level Video Understanding in the Wild Challenge.