# Qwen3-VL-Seg: Unlocking Open-World Referring Segmentation with Vision-Language Grounding

arXiv: https://arxiv.org/abs/2605.07141 · HTML: https://mlsanity.com/paper/2605.07141/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 34 of its 39 references that are in the map.
- First posted on arXiv 8 May 2026.

## Builds on (most cited first, 30 of 34)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [GPT-4o System Card](https://mlsanity.com/paper/2410.21276/) (2024)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://mlsanity.com/paper/2303.05499/) (2023)
- [SAM 2: Segment Anything in Images and Videos](https://mlsanity.com/paper/2408.00714/) (2024)
- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency](https://mlsanity.com/paper/2508.18265/) (2025)
- [Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression](https://mlsanity.com/paper/1902.09630/) (2019)
- [LVIS: A Dataset for Large Vocabulary Instance Segmentation](https://mlsanity.com/paper/1908.03195/) (2019)
- [Modeling Context in Referring Expressions](https://mlsanity.com/paper/1608.00272/) (2016)
- [OpenAI GPT-5 System Card](https://mlsanity.com/paper/2601.03267/) (2025)
- [Shikra: Unleashing Multimodal LLM's Referential Dialogue Magic](https://mlsanity.com/paper/2306.15195/) (2023)
- [SAM 3: Segment Anything with Concepts](https://mlsanity.com/paper/2511.16719/) (2025)
- [Generation and Comprehension of Unambiguous Object Descriptions](https://mlsanity.com/paper/1511.02283/) (2015)
- [LISA: Reasoning Segmentation via Large Language Model](https://mlsanity.com/paper/2308.00692/) (2023)
- [Ferret: Refer and Ground Anything Anywhere at Any Granularity](https://mlsanity.com/paper/2310.07704/) (2023)
- [Florence-2: Advancing a Unified Representation for a Variety of Vision Tasks](https://mlsanity.com/paper/2311.06242/) (2023)
- [Recognize Anything: A Strong Image Tagging Model](https://mlsanity.com/paper/2306.03514/) (2023)
- [Segmentation from Natural Language Expressions](https://mlsanity.com/paper/1603.06180/) (2016)
- [GRES: Generalized Referring Expression Segmentation](https://mlsanity.com/paper/2306.00968/) (2023)
- [PixelLM: Pixel Reasoning with Large Multimodal Model](https://mlsanity.com/paper/2312.02228/) (2023)
- [GSVA: Generalized Segmentation via Multimodal Large Language Models](https://mlsanity.com/paper/2312.10103/) (2023)
- [VisionLLM v2: An End-to-End Generalist Multimodal Large Language Model for Hundreds of Vision-Language Tasks](https://mlsanity.com/paper/2406.08394/) (2024)
- [SAM4MLLM: Enhance Multi-Modal Large Language Model for Referring Expression Segmentation](https://mlsanity.com/paper/2409.10542/) (2024)
- [Text4Seg: Reimagining Image Segmentation as Text Generation](https://mlsanity.com/paper/2410.09855/) (2024)
- [PerceptionGPT: Effectively Fusing Visual Perception into LLM](https://mlsanity.com/paper/2311.06612/) (2023)
- [Multi-modal Instruction Tuned LLMs with Fine-grained Visual Perception](https://mlsanity.com/paper/2403.02969/) (2024)
- [SegAgent: Exploring Pixel Understanding Capabilities in MLLMs by Imitating Human Annotator Trajectories](https://mlsanity.com/paper/2503.08625/) (2025)

## Built on by (most cited first, 5 of 5)

- [Personalized Object Identification and Localization via In-Context Inference with Vision-Language Models](https://mlsanity.com/paper/2607.00357/) (2026)
- [PixVL: Self-Supervised Training of Pixel-Level MLLMs via a Unified Mask--Text Consistency Cycle](https://mlsanity.com/paper/2608.01354/) (2026)
- [Vision-Language Grounding as Bidirectional Concept Correspondence](https://mlsanity.com/paper/2608.07886/) (2026)
- [TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents](https://mlsanity.com/paper/2608.12898/) (2026)
- [Groundbench: Multi-Resolution Polygon Grounding Exposes the Geometry Gap in Vision-Language Models](https://mlsanity.com/paper/2609.27821/) (2026)

## Abstract (arXiv)

Open-world referring segmentation requires grounding unconstrained language expressions to precise pixel-level regions. Existing multimodal large language models (MLLMs) exhibit strong open-world visual grounding, but their outputs remain limited to sparse bounding-box coordinates and are insufficient for dense visual prediction. Recent MLLM-based segmentation methods either directly predict sparse contour coordinates, struggling to reconstruct continuous object boundaries, or rely on external segmentation foundation models such as the Segment Anything Model (SAM), introducing substantial architectural and deployment overhead. We present Qwen3-VL-Seg, a parameter-efficient framework that treats the MLLM-predicted box as a semantically grounded structural prior and decodes it into pixel-level referring segmentation. At its core, a lightweight box-guided mask decoder combines multi-scale spatial feature injection, spatial-semantic query construction, box-guided high-resolution pixel fusion, and iterative mask-aware query refinement, introducing only 17M parameters (about 0.4% of the base model). For scalable open-world training, we construct SA1B-ORS, an SA-1B-derived dataset with two subsets: SA1B-CoRS (category-oriented samples) and SA1B-DeRS (descriptive, instance-specific samples). For evaluation, we curate ORS-Bench, a manually screened benchmark with in-distribution and out-of-distribution subsets covering diverse referring expression types. Extensive experiments on referring expression segmentation, visual grounding, and ORS-Bench show that Qwen3-VL-Seg performs strongly across closed-set and open-world settings, with clear advantages on language-intensive instructions and strong out-of-distribution generalization. Evaluations on general multimodal benchmarks further show that the model broadly preserves general-purpose multimodal competence after segmentation-oriented adaptation.