MIRG-RL: Multi-Image Reasoning and Grounding with Reinforcement Learning

• URL: http://arxiv.org/abs/2509.21788v1
• Date: Fri, 26 Sep 2025 02:43:22 GMT
• Title: MIRG-RL: Multi-Image Reasoning and Grounding with Reinforcement Learning
• Abstract summary: Current Large Visual Language Models (LVLMs) face two critical challenges: the lack of cross-image reasoning capabilities and insufficient cross-image reference reward modeling.<n>We propose a unified framework - Multi-Image Reasoning and Grounding with Reinforcement Learning (MIRG-RL)<n>Specifically, our two-stage training paradigm combines supervised fine-tuning with annotated trajectories and image-aware reinforcement learning optimization.
• Score: 10.049259114211663
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Multi-image reasoning and grounding require understanding complex cross-image relationships at both object levels and image levels. Current Large Visual Language Models (LVLMs) face two critical challenges: the lack of cross-image reasoning capabilities and insufficient cross-image reference reward modeling. To address these issues, we propose a unified framework - Multi-Image Reasoning and Grounding with Reinforcement Learning (MIRG-RL). Specifically, our two-stage training paradigm combines supervised fine-tuning with annotated trajectories and image-aware reinforcement learning optimization, progressively developing multi-image reasoning capabilities. Furthermore, we innovatively propose a method for constructing the trajectory data, which integrates object-level and image-level annotation information, and use this method to generate a lightweight reasoning-enhanced dataset. To effectively resolve cross-image ambiguities, we design an image-aware RL policy with dual reward functions for objects and images. Experiments demonstrate that MIRG-RL achieves state-of-the-art (SOTA) performance in multi-image grounding benchmarks, attaining 64.82% on cross-image reasoning tasks - exceeding the previous best method by 1%. The code and dataset have been released at https://github.com/ZEUS2035/MIRG-RL.

Related papers

• RL-RIG: A Generative Spatial Reasoner via Intrinsic Reflection [18.52946282633359]
  RL-RIG is a Reinforcement Learning framework for Reflection-based Image Generation.<n>We develop Reflection-GRPO to train the VLM Actor for edit prompts and the Image Editor for better image quality under a given prompt.<n> Experimental results show that RL-RIG outperforms existing state-of-the-art open-source models by up to 11% in terms of controllable and precise spatial reasoning in image generation.
  arXiv  Detail & Related papers  (2026-02-23T15:39:53Z)
• More Images, More Problems? A Controlled Analysis of VLM Failure Modes [80.64323947730905]
  Large Vision Language Models (LVLMs) have demonstrated remarkable capabilities, yet their proficiency in understanding and reasoning over multiple images remains largely unexplored.<n>We introduce MIMIC, a new benchmark designed to rigorously evaluate the multi-image capabilities of LVLMs.
  arXiv  Detail & Related papers  (2026-01-12T18:45:13Z)
• GeM-VG: Towards Generalized Multi-image Visual Grounding with Multimodal Large Language Models [30.759062684007873]
  Multimodal Large Language Models (MLLMs) have demonstrated impressive progress in single-image grounding and general multi-image understanding.<n>We propose GeM-VG, an MLLM capable of Generalized Multi-image Visual Grounding.
  arXiv  Detail & Related papers  (2026-01-08T09:58:35Z)
• Training Multi-Image Vision Agents via End2End Reinforcement Learning [51.81337984526068]
  We propose IMAgent, an open-source vision agent trained via end-to-end reinforcement learning.<n>By leveraging a multi-agent system, we generate challenging and visually-rich multi-image QA pairs.<n>We develop two specialized tools for visual reflection and confirmation, allowing the model to proactively reallocate its attention to image content.
  arXiv  Detail & Related papers  (2025-12-05T10:02:38Z)
• PeRL: Permutation-Enhanced Reinforcement Learning for Interleaved Vision-Language Reasoning [50.21619363035618]
  We propose a general reinforcement learning approach PeRL tailored for interleaved multimodal tasks.<n>We introduce permutation of image sequences to simulate varied positional relationships to explore more spatial and positional diversity.<n>Our experiments confirm that PeRL trained model consistently surpasses R1-related and interleaved VLM baselines by a large margin.
  arXiv  Detail & Related papers  (2025-06-17T18:25:56Z)
• Migician: Revealing the Magic of Free-Form Multi-Image Grounding in Multimodal Large Language Models [79.59567114769513]
  We introduce Migician, the first multi-image grounding model capable of performing free-form and accurate grounding across multiple images.<n>Our model achieves significantly superior multi-image grounding capabilities, outperforming the best existing MLLMs by 24.94% and even surpassing much larger 70B models.
  arXiv  Detail & Related papers  (2025-01-10T07:56:23Z)
• RL-I2IT: Image-to-Image Translation with Deep Reinforcement Learning [54.40719981158774]
  Image-to-Image Translation (I2IT) methods generate images in a single run of a deep learning (DL) model. We reformulate I2IT as a step-wise decision-making problem via deep reinforcement learning (DRL) We propose a novel framework that performs RL-based I2IT (RL-I2IT)
  arXiv  Detail & Related papers  (2023-09-24T15:40:40Z)
• Improving Human-Object Interaction Detection via Virtual Image Learning [68.56682347374422]
  Human-Object Interaction (HOI) detection aims to understand the interactions between humans and objects. In this paper, we propose to alleviate the impact of such an unbalanced distribution via Virtual Image Leaning (VIL) A novel label-to-image approach, Multiple Steps Image Creation (MUSIC), is proposed to create a high-quality dataset that has a consistent distribution with real images.
  arXiv  Detail & Related papers  (2023-08-04T10:28:48Z)
• DDet: Dual-path Dynamic Enhancement Network for Real-World Image Super-Resolution [69.2432352477966]
  Real image super-resolution(Real-SR) focus on the relationship between real-world high-resolution(HR) and low-resolution(LR) image. In this article, we propose a Dual-path Dynamic Enhancement Network(DDet) for Real-SR. Unlike conventional methods which stack up massive convolutional blocks for feature representation, we introduce a content-aware framework to study non-inherently aligned image pair.
  arXiv  Detail & Related papers  (2020-02-25T18:24:51Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.