Research map / Video / Video-language understanding
InstructSeg: Unifying Instructed Visual Segmentation with Multi-modal Large Language Models
Cong Wei, Yujie Zhong, Haoxian Tan, Yingsen Zeng, Yong Liu, Hongfa Wang, Yujiu Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add OVP and VMTF together (ablation) | InstructSeg · RefCOCO val | cIoU: 83.3 -> 85.8 | Improved |
| add OVP and VMTF together (ablation) | InstructSeg · ReVOS overall | J&F: 52.9 -> 54.5 | Improved |
| add Object-aware Video Perceiver (ablation) | InstructSeg without modules · ReVOS overall | J&F: 52.9 -> 54.0 | Improved |
| add Vision-guided Multi-granularity Text Fusion (ablation) | InstructSeg without modules · RefCOCO val | cIoU: 83.3 -> 84.8 | Improved |
| increase reference frames from 0 to 4 (ablation) | InstructSeg OVP · ReVOS overall | J&F: 53.6 -> 54.5 | Improved |
| increase reference frames from 4 to 8 (ablation) | InstructSeg OVP · ReVOS overall | J&F: 54.5 -> 54.6 | No clear effect |
| fuse detailed text with global text instead of global only (ablation) | InstructSeg VMTF · RefCOCO val | cIoU: 84.2 -> 85.8 | Improved |
| fuse global text with detailed text instead of detailed only (ablation) | InstructSeg VMTF · ReVOS reasoning | J&F: 51.6 -> 51.9 | Improved |
| add video training data to image-only training (ablation) | InstructSeg · ReVOS overall | J&F: 48.7 -> 54.5 | Improved |
| add image training data to video-only training (ablation) | InstructSeg · RefCOCO val | cIoU: 78.1 -> 85.8 | Improved |
| add referring data to reasoning-only training (ablation) | InstructSeg · RefCOCO val | cIoU: 74.0 -> 85.8 | Improved |
| add reasoning data to referring-only training (ablation) | InstructSeg · ReVOS reasoning | J&F: 37.4 -> 51.9 | Improved |
Later papers that compared with this method
2 later accepted papers report InstructSeg as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- VIRST: Video-Instructed Reasoning Assistant for SpatioTemporal Segmentation (CVPR 2026 · better on 1 of 2 shared results)
- Training-Free Spatio-temporal Decoupled Reasoning Video Segmentation with Adaptive Object Memory (AAAI 2026 · better on 2 of 2 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- VISA-7B — VISA: Reasoning Video Object Segmentation via Large Language Model (ECCV 2024) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- One Token to Seg Them All: Language Instructed Reasoning Segmentation in Videos (NeurIPS 2024 · similarity 0.90)
- HyperSeg: Hybrid Segmentation Assistant with Fine-grained Visual Perceiver (CVPR 2025 · similarity 0.90)
- Seg-ReSearch: Segmentation with Interleaved Reasoning and External Search (ICML 2026 · similarity 0.89)
- LISA: Reasoning Segmentation via Large Language Model (CVPR 2024 · similarity 0.88)
- UniPixel: Unified Object Referring and Segmentation for Pixel-Level Visual Reasoning (NeurIPS 2025 · similarity 0.88)
- Hugging Visual Prompt and Segmentation Tokens: Consistency Learning for Fine-Grained Visual Understanding in MLLMs (CVPR 2026 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
