Learning to Focus and Precise Cropping: A Reinforcement Learning Framework with Information Gaps and Grounding Loss for MLLMs

TL;DR

This paper introduces a two-stage reinforcement learning framework that improves multimodal large language models' ability to focus on relevant image regions for better question answering, using information gaps and grounding loss.

Contribution

It presents a novel RL-based approach that enhances cropping precision and attention in MLLMs without requiring trajectory supervision, achieving state-of-the-art results.

Findings

Significantly improves attention to cropped regions in MLLMs.

Achieves state-of-the-art performance on high-resolution visual question-answering benchmarks.

Efficiently perceives and reasons about fine-grained visual details.

Abstract

To enhance the perception and reasoning capabilities of multimodal large language models in complex visual scenes, recent research has introduced agent-based workflows. In these works, MLLMs autonomously utilize image cropping tool to analyze regions of interest for question answering. While existing training strategies, such as those employing supervised fine-tuning and reinforcement learning, have made significant progress, our empirical analysis reveals a key limitation. We demonstrate the model's strong reliance on global input and its weak dependence on the details within the cropped region. To address this issue, we propose a novel two-stage reinforcement learning framework that does not require trajectory supervision. In the first stage, we introduce the ``Information Gap" mechanism by adjusting the granularity of the global image. This mechanism trains the model to answer questions by focusing on cropped key regions, driven by the information gain these regions provide. The second stage further enhances cropping precision by incorporating a grounding loss, using a small number of bounding box annotations. Experiments show that our method significantly enhances the model's attention to cropped regions, enabling it to achieve state-of-the-art performance on high-resolution visual question-answering benchmarks. Our method provides a more efficient approach for perceiving and reasoning fine-grained details in MLLMs. Code is available at: https://github.com/XuanPu-Z/LFPC.