mlsanityData to 7 Oct 2026

The 1st Solution for 4th PVUW MeViS Challenge: Unleashing the Potential of Large Multimodal Models for Referring Video Segmentation

arXiv 2504.05178 ↗First posted 7 Apr 2025cs.CVcited by 3

On the map: Multimodal LLMs / vision-language models → VLM benchmarks and visual reasoning

Abstract (arXiv)

Motion expression video segmentation is designed to segment objects in accordance with the input motion expressions. In contrast to the conventional Referring Video Object Segmentation (RVOS), it places emphasis on motion as well as multi-object expressions, making it more arduous. Recently, Large Multimodal Models (LMMs) have begun to shine in RVOS due to their powerful vision-language perception capabilities. In this work, we propose a simple and effective inference optimization method to fully unleash the potential of LMMs in referring video segmentation. Firstly, we use Sa2VA as our baseline, which is a unified LMM for dense grounded understanding of both images and videos. Secondly, we uniformly sample the video frames during the inference process to enhance the model's understanding of the entire video. Finally, we integrate the results of multiple expert models to mitigate the erroneous predictions of a single model. Our solution achieved 61.98% J&F on the MeViS test set and ranked 1st place in the 4th PVUW Challenge MeViS Track at CVPR 2025.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 25 of its 27 references that are in the map. First posted on arXiv 7 Apr 2025.

Builds on

25 of 27 references.

Closest ancestors · 3

Same direction, newest first

  1. 2025Sa2VA: Marrying SAM2 with MLLM for Dense Grounded Understanding of Images and Videos7 Jan
  2. 2023PixelLM: Pixel Reasoning with Large Multimodal Model4 Dec
  3. 2023LISA: Reasoning Segmentation via Large Language Model1 Aug

Other references · 22

  1. 2023Visual Instruction Tuningcited by 7,792
  2. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 4,151
  3. 2024SAM 2: Segment Anything in Images and Videoscited by 2,784
  4. 2024Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scalingcited by 1,573
  5. 2023MOSE: A New Dataset for Video Object Segmentation in Complex Scenescited by 232
  6. 2023MeViS: A Large-scale Benchmark for Video Segmentation with Motion Expressionscited by 225
  7. 2023Universal Instance Perception as Object Discovery and Retrievalcited by 190
  8. 2023Putting the Object Back into Video Object Segmentationcited by 182
  9. 2022Language as Queries for Referring Video Object Segmentationcited by 180
  10. 2018Video Object Segmentation with Language Referring Expressionscited by 161
  11. 2021End-to-End Referring Video Object Segmentation with Multimodal Transformerscited by 135
  12. 2024VISA: Reasoning Video Object Segmentation via Large Language Modelscited by 123
10 more
  1. 2022VITA: Video Instance Segmentation via Object Token Associationcited by 94
  2. 2023SOC: Semantic-Assisted Object Cluster for Referring Video Object Segmentationcited by 57
  3. 2024Decoupling Static and Hierarchical Motion Perception for Referring Video Segmentationcited by 55
  4. 2023Referred by Multi-Modality: A Unified Temporal Transformer for Video Object Segmentationcited by 54
  5. 2021Rethinking Cross-modal Interaction from a Top-down Perspective for Referring Video Object Segmentationcited by 42
  6. 2024Unified Embedding Alignment for Open-Vocabulary Video Instance Segmentationcited by 15
  7. 2024PVUW 2024 Challenge on Complex Video Understanding: Methods and Resultscited by 12
  8. 2024LSVOS Challenge Report: Large-scale Complex and Long Video Object Segmentationcited by 8
  9. 2024Video Object Segmentation via SAM 2: The 4th Solution for LSVOS Challenge VOS Trackcited by 3
  10. 2024UNINEXT-Cutie: The 1st Solution for LSVOS Challenge RVOS Trackcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 2; 2026: 1.

3 papers.

  1. 2025The 1st Solution for 7th LSVOS RVOS Track: SaSaSa2VAcited by 7
  2. 2025LSVOS 2025 Challenge Report: Recent Advances in Complex Video Object Segmentationcited by 4
  3. 2026Mathematical Reasoning via Intervention-Based Time-Series Causal Discovery Using LLMs as Concept Mastery Simulatorscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 20254th PVUW MeViS 3rd Place Report: Sa2VA
  2. 2026The 1st Winner for 5th PVUW MeViS-Text Challenge: Strong MLLMs Meet SAM3 for Referring Video Object Segmentation
  3. 2025Enhancing Sa2VA for Referent Video Object Segmentation: 2nd Solution for 7th LSVOS RVOS Track
  4. 2026Report of the 8th LSVOS Challenge: Complex and Multimodal Video Object Segmentation
  5. 20243rd Place Solution for MeViS Track in CVPR 2024 PVUW workshop: Motion Expression guided Video Segmentation
  6. 2026Weakly-Supervised Referring Video Object Segmentation through Text Supervision
  7. 2025Unleashing Hierarchical Reasoning: An LLM-Driven Framework for Training-Free Referring Video Object Segmentation
  8. 2025ReferDINO: Referring Video Object Segmentation with Visual Grounding Foundations