# LENS: Learning to Segment Anything with Unified Reinforced Reasoning

arXiv: https://arxiv.org/abs/2508.14153 · HTML: https://mlsanity.com/paper/2508.14153/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 18 of its 46 references that are in the map.
- First posted on arXiv 19 Aug 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)

## Builds on (most cited first, 18 of 18)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Constitutional AI: Harmlessness from AI Feedback](https://mlsanity.com/paper/2212.08073/) (2022)
- [SAM 2: Segment Anything in Images and Videos](https://mlsanity.com/paper/2408.00714/) (2024)
- [Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond](https://mlsanity.com/paper/2308.12966/) (2023)
- [Decoupled Weight Decay Regularization](https://mlsanity.com/paper/1711.05101/) (2017)
- [Large Language Models Can Self-Improve](https://mlsanity.com/paper/2210.11610/) (2022)
- [VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model](https://mlsanity.com/paper/2504.07615/) (2025)
- [Personalize Segment Anything Model with One Shot](https://mlsanity.com/paper/2305.03048/) (2023)
- [Seg-Zero: Reasoning-Chain Guided Segmentation via Cognitive Reinforcement](https://mlsanity.com/paper/2503.06520/) (2025)
- [Perception-R1: Pioneering Perception Policy with Reinforcement Learning](https://mlsanity.com/paper/2504.07954/) (2025)
- [EVF-SAM: Early Vision-Language Fusion for Text-Prompted Segment Anything Model](https://mlsanity.com/paper/2406.20076/) (2024)
- [HyperSeg: Towards Universal Visual Segmentation with Large Language Model](https://mlsanity.com/paper/2411.17606/) (2024)
- [InstructSeg: Unifying Instructed Visual Segmentation with Multi-modal Large Language Models](https://mlsanity.com/paper/2412.14006/) (2024)
- [GroundingSuite: Measuring Complex Multi-Granular Pixel Grounding](https://mlsanity.com/paper/2503.10596/) (2025)

## Built on by (most cited first, 18 of 18)

- [SAMTok: Representing Any Mask with Two Words](https://mlsanity.com/paper/2601.16093/) (2026)
- [Vision as Unified Multimodal Generation](https://mlsanity.com/paper/2607.06560/) (2026)
- [Evolving, Not Training: Zero-Shot Reasoning Segmentation via Evolutionary Prompting](https://mlsanity.com/paper/2512.24702/) (2025)
- [DR2Seg: Decomposed Two-Stage Rollouts for Efficient Reasoning Segmentation in Multimodal Large Language Models](https://mlsanity.com/paper/2601.09981/) (2026)
- [Reason Twice: Segmentation via Candidate Discovery and Comparative Reasoning](https://mlsanity.com/paper/2606.09303/) (2026)
- [CORE-Seg: Reasoning-Driven Segmentation for Complex Lesions via Reinforcement Learning](https://mlsanity.com/paper/2603.05911/) (2026)
- [Topo-R1: Detecting Topological Anomalies via Vision-Language Models](https://mlsanity.com/paper/2603.13054/) (2026)
- [Tarot-SAM3: Training-free SAM3 for Any Referring Expression Segmentation](https://mlsanity.com/paper/2604.07916/) (2026)
- [Vision Harnessing Agent for Open Ad-hoc Segmentation](https://mlsanity.com/paper/2605.19410/) (2026)
- [B-GRTO: Bootstrapped Group Relative Tool Optimization for Referring Segmentation](https://mlsanity.com/paper/2605.23500/) (2026)
- [Actor as Its Own Critic: Unifying Region Understanding and Localization via CycleGRPO](https://mlsanity.com/paper/2607.11581/) (2026)
- [Mixture-of-Thought-Tokens: Unifying Perception and Reasoning for Free-form Multimodal Grounding](https://mlsanity.com/paper/2607.24407/) (2026)
- [PinPoint: Prompting with Informative Interior Points](https://mlsanity.com/paper/2605.26689/) (2026)
- [VideoSEG-O3: A Multi-turn Reinforcement Learning Framework for Reasoning Video Object Segmentation](https://mlsanity.com/paper/2606.06819/) (2026)
- [CFCamo: A Counterfactual Detect-or-Abstain Framework for Camouflaged Object Detection](https://mlsanity.com/paper/2606.11231/) (2026)
- [DGSeg: Dynamic Gating of Semantic-Spatial Guided Predictions for Reasoning Segmentation](https://mlsanity.com/paper/2607.04779/) (2026)
- [EgoAfford: Task-Oriented Affordance Grounding via Egocentric Referring Segmentation](https://mlsanity.com/paper/2608.04533/) (2026)
- [CoEvolve: Construct-to-Edit Visual Grounding with Bidirectional State Refinement](https://mlsanity.com/paper/2610.01710/) (2026)

## Abstract (arXiv)

Text-prompted image segmentation enables fine-grained visual understanding and is critical for applications such as human-computer interaction and robotics. However, existing supervised fine-tuning methods typically ignore explicit chain-of-thought (CoT) reasoning at test time, which limits their ability to generalize to unseen prompts and domains. To address this issue, we introduce LENS, a scalable reinforcement-learning framework that jointly optimizes the reasoning process and segmentation in an end-to-end manner. We propose unified reinforcement-learning rewards that span sentence-, box-, and segment-level cues, encouraging the model to generate informative CoT rationales while refining mask quality. Using a publicly available 3-billion-parameter vision-language model, i.e., Qwen2.5-VL-3B-Instruct, LENS achieves an average cIoU of 81.2% on the RefCOCO, RefCOCO+, and RefCOCOg benchmarks, outperforming the strong fine-tuned method, i.e., GLaMM, by up to 5.6%. These results demonstrate that RL-driven CoT reasoning significantly enhances text-prompted segmentation and offers a practical path toward more generalizable Segment Anything models (SAM). Code is available at https://github.com/hustvl/LENS.