Active-O3: Empowering Multimodal Large Language Models with Active Perception via GRPO

TL;DR

This paper introduces ACTIVE-O3, a reinforcement learning framework that enhances multimodal large language models with active perception capabilities, improving efficiency and accuracy in various perception tasks.

Contribution

We propose ACTIVE-O3, a novel RL-based training framework built on GRPO to equip MLLMs with active perception, along with a comprehensive benchmark suite for evaluation.

Findings

ACTIVE-O3 improves search efficiency and region selection accuracy.

It demonstrates strong zero-shot reasoning abilities on the V* Benchmark.

The framework performs well across diverse perception tasks.

Abstract

Active vision, also known as active perception, refers to the process of actively selecting where and how to look in order to gather task-relevant information. It is a critical component of efficient perception and decision-making in humans and advanced embodied agents. Recently, the use of Multimodal Large Language Models (MLLMs) as central planning and decision-making modules in robotic systems has gained extensive attention. However, despite the importance of active perception in embodied intelligence, there is little to no exploration of how MLLMs can be equipped with or learn active perception capabilities. In this paper, we first provide a systematic definition of MLLM-based active perception tasks. We point out that the recently proposed GPT-o3 model's zoom-in search strategy can be regarded as a special case of active perception; however, it still suffers from low search efficiency and inaccurate region selection. To address these issues, we propose ACTIVE-O3, a purely reinforcement learning based training framework built on top of GRPO, designed to equip MLLMs with active perception capabilities. We further establish a comprehensive benchmark suite to evaluate ACTIVE-O3 across both general open-world tasks, such as small-object and dense object grounding, and domain-specific scenarios, including small object detection in remote sensing and autonomous driving, as well as fine-grained interactive segmentation. In addition, ACTIVE-O3 also demonstrates strong zero-shot reasoning abilities on the V* Benchmark, without relying on any explicit reasoning data. We hope that our work can provide a simple codebase and evaluation protocol to facilitate future research on active perception in MLLMs.