FVOS for MOSE Track of 4th PVUW Challenge: 3rd Place Solution

2025-04-13

AI Summary

To address low segmentation accuracy and severe inter-frame object fragmentation in video object segmentation (VOS) under complex real-world scenarios, this work proposes the FVOS fine-tuning framework. First, a lightweight domain-adaptive fine-tuning is applied to a pre-trained model. Second, a morphology-based post-processing module—leveraging morphological opening and closing operations—is introduced to explicitly repair intra-frame target discontinuities. Third, a multi-scale feature extraction mechanism coupled with weighted voting fusion is designed to enhance temporal consistency and robustness. Evaluated on the PVUW 2025 MOSE challenge, our method achieves J&F scores of 76.81% on the validation set and 83.92% on the test set, ranking third overall. The framework significantly mitigates segmentation degradation caused by dynamic backgrounds, occlusions, and small objects.

Technology Category

• Computer Vision: Segmentation
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Video Object Segmentation (VOS) is one of the most fundamental and challenging tasks in computer vision and has a wide range of applications. Most existing methods rely on spatiotemporal memory networks to extract frame-level features and have achieved promising results on commonly used datasets. However, these methods often struggle in more complex real-world scenarios. This paper addresses this issue, aiming to achieve accurate segmentation of video objects in challenging scenes. We propose fine-tuning VOS (FVOS), optimizing existing methods for specific datasets through tailored training. Additionally, we introduce a morphological post-processing strategy to address the issue of excessively large gaps between adjacent objects in single-model predictions. Finally, we apply a voting-based fusion method on multi-scale segmentation results to generate the final output. Our approach achieves J&F scores of 76.81% and 83.92% during the validation and testing stages, respectively, securing third place overall in the MOSE Track of the 4th PVUW challenge 2025.

Problem

• Improving video object segmentation in complex real-world scenarios
• Optimizing VOS methods via fine-tuning for specific datasets
• Reducing gaps between objects using morphological post-processing

Innovation

• Fine-tuning VOS for specific datasets optimization
• Morphological post-processing for gap reduction
• Voting-based fusion on multi-scale results