SAM-R1: Leveraging SAM for Reward Feedback in Multimodal Segmentation via Reinforcement Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Jiaqi, Xu, Zunnan, Zhou, Jun, Liu, Ting, Xiao, Yicheng, Ou, Mingwen, Ji, Bowen, Li, Xiu, Yuan, Kehong
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911516605808640
author Huang, Jiaqi
Xu, Zunnan
Zhou, Jun
Liu, Ting
Xiao, Yicheng
Ou, Mingwen
Ji, Bowen
Li, Xiu
Yuan, Kehong
author_facet Huang, Jiaqi
Xu, Zunnan
Zhou, Jun
Liu, Ting
Xiao, Yicheng
Ou, Mingwen
Ji, Bowen
Li, Xiu
Yuan, Kehong
contents Leveraging multimodal large models for image segmentation has become a prominent research direction. However, existing approaches typically rely heavily on manually annotated datasets that include explicit reasoning processes, which are costly and time-consuming to produce. Recent advances suggest that reinforcement learning (RL) can endow large models with reasoning capabilities without requiring such reasoning-annotated data. In this paper, we propose SAM-R1, a novel framework that enables multimodal large models to perform fine-grained reasoning in image understanding tasks. Our approach is the first to incorporate fine-grained segmentation settings during the training of multimodal reasoning models. By integrating task-specific, fine-grained rewards with a tailored optimization objective, we further enhance the model's reasoning and segmentation alignment. We also leverage the Segment Anything Model (SAM) as a strong and flexible reward provider to guide the learning process. With only 3k training samples, SAM-R1 achieves strong performance across multiple benchmarks, demonstrating the effectiveness of reinforcement learning in equipping multimodal models with segmentation-oriented reasoning capabilities.
format Preprint
id arxiv_https___arxiv_org_abs_2505_22596
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SAM-R1: Leveraging SAM for Reward Feedback in Multimodal Segmentation via Reinforcement Learning
Huang, Jiaqi
Xu, Zunnan
Zhou, Jun
Liu, Ting
Xiao, Yicheng
Ou, Mingwen
Ji, Bowen
Li, Xiu
Yuan, Kehong
Computer Vision and Pattern Recognition
Leveraging multimodal large models for image segmentation has become a prominent research direction. However, existing approaches typically rely heavily on manually annotated datasets that include explicit reasoning processes, which are costly and time-consuming to produce. Recent advances suggest that reinforcement learning (RL) can endow large models with reasoning capabilities without requiring such reasoning-annotated data. In this paper, we propose SAM-R1, a novel framework that enables multimodal large models to perform fine-grained reasoning in image understanding tasks. Our approach is the first to incorporate fine-grained segmentation settings during the training of multimodal reasoning models. By integrating task-specific, fine-grained rewards with a tailored optimization objective, we further enhance the model's reasoning and segmentation alignment. We also leverage the Segment Anything Model (SAM) as a strong and flexible reward provider to guide the learning process. With only 3k training samples, SAM-R1 achieves strong performance across multiple benchmarks, demonstrating the effectiveness of reinforcement learning in equipping multimodal models with segmentation-oriented reasoning capabilities.
title SAM-R1: Leveraging SAM for Reward Feedback in Multimodal Segmentation via Reinforcement Learning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2505.22596