VisionReasoner: Unified Reasoning-Integrated Visual Perception via Reinforcement Learning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917255836598272 |
|---|---|
| author | Liu, Yuqi Qu, Tianyuan Zhong, Zhisheng Peng, Bohao Liu, Shu Yu, Bei Jia, Jiaya |
| author_facet | Liu, Yuqi Qu, Tianyuan Zhong, Zhisheng Peng, Bohao Liu, Shu Yu, Bei Jia, Jiaya |
| contents | Large vision-language models exhibit inherent capabilities to handle diverse visual perception tasks. In this paper, we introduce VisionReasoner, a unified framework capable of reasoning and solving multiple visual perception tasks within a shared model. Specifically, by designing a unified reward mechanism and multi-object cognitive learning strategies, VisionReasoner enhances its reasoning capabilities to analyze visual inputs, and addresses diverse perception tasks within a unified model. VisionReasoner generates a structured reasoning process before delivering the desired outputs responding to user queries. Human evaluation reveals the reasoning process of VisionReasoner is faithful and reliable even without annotated reasoning train data. To rigorously assess unified visual perception capabilities, we evaluate VisionReasoner on ten diverse tasks spanning three critical domains: detection, segmentation, and counting. Experimental results show that VisionReasoner achieves superior performance as a unified model, outperforming the baseline Qwen2.5VL by relative margins of 29.1\% on COCO (detection), 22.1\% on ReasonSeg (segmentation), and 13.2\% on CountBench (counting). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_12081 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | VisionReasoner: Unified Reasoning-Integrated Visual Perception via Reinforcement Learning Liu, Yuqi Qu, Tianyuan Zhong, Zhisheng Peng, Bohao Liu, Shu Yu, Bei Jia, Jiaya Computer Vision and Pattern Recognition Large vision-language models exhibit inherent capabilities to handle diverse visual perception tasks. In this paper, we introduce VisionReasoner, a unified framework capable of reasoning and solving multiple visual perception tasks within a shared model. Specifically, by designing a unified reward mechanism and multi-object cognitive learning strategies, VisionReasoner enhances its reasoning capabilities to analyze visual inputs, and addresses diverse perception tasks within a unified model. VisionReasoner generates a structured reasoning process before delivering the desired outputs responding to user queries. Human evaluation reveals the reasoning process of VisionReasoner is faithful and reliable even without annotated reasoning train data. To rigorously assess unified visual perception capabilities, we evaluate VisionReasoner on ten diverse tasks spanning three critical domains: detection, segmentation, and counting. Experimental results show that VisionReasoner achieves superior performance as a unified model, outperforming the baseline Qwen2.5VL by relative margins of 29.1\% on COCO (detection), 22.1\% on ReasonSeg (segmentation), and 13.2\% on CountBench (counting). |
| title | VisionReasoner: Unified Reasoning-Integrated Visual Perception via Reinforcement Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2505.12081 |