LSVOS 2025 Challenge Report: Recent Advances in Complex Video Object Segmentation

2025-10-13

AI Summary

Complex video object segmentation (VOS) faces core challenges including poor long-term consistency and weak generalization under dense small objects, frequent object disappearance/reappearance, severe occlusion, and adverse conditions (e.g., rain, fog, low illumination). To address these, we introduce MOSEv2—the first VOS benchmark targeting real-world complex scenarios—and propose an enhanced J&F metric emphasizing multi-scale and disappeared-object evaluation. Methodologically, we design a language-vision collaborative architecture that integrates multimodal large model (MLLM)-driven semantic understanding with memory-aware propagation, significantly improving language-guided accuracy and temporal consistency modeling. Experiments demonstrate state-of-the-art performance on MOSEv2. Furthermore, we systematically analyze technical evolution trends toward open-world robust VOS, establishing MOSEv2 as a new standard benchmark and providing concrete directions for future research.

Technology Category

• Computer Vision: Language and Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• 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
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

This report presents an overview of the 7th Large-scale Video Object Segmentation (LSVOS) Challenge held in conjunction with ICCV 2025. Besides the two traditional tracks of LSVOS that jointly target robustness in realistic video scenarios: Classic VOS (VOS), and Referring VOS (RVOS), the 2025 edition features a newly introduced track, Complex VOS (MOSEv2). Building upon prior insights, MOSEv2 substantially increases difficulty, introducing more challenging but realistic scenarios including denser small objects, frequent disappear/reappear events, severe occlusions, adverse weather and lighting, etc., pushing long-term consistency and generalization beyond curated benchmarks. The challenge retains standard ${J}$, $F$, and ${J&F}$ metrics for VOS and RVOS, while MOSEv2 adopts ${J&dot{F}}$ as the primary ranking metric to better evaluate objects across scales and disappearance cases. We summarize datasets and protocols, highlight top-performing solutions, and distill emerging trends, such as the growing role of LLM/MLLM components and memory-aware propagation, aiming to chart future directions for resilient, language-aware video segmentation in the wild.

Problem

• Advancing video object segmentation in complex realistic scenarios
• Addressing long-term consistency under occlusion and disappearance
• Improving generalization beyond curated benchmarks through new metrics

Innovation

• Introduces Complex VOS track with realistic scenarios
• Adopts J&F metric for multi-scale object evaluation
• Integrates LLM components for language-aware video segmentation