paper-with-me

Papers

Introducing Visual Perception Token into Multimodal Large Language Model

2025-02-24 · Runpeng Yu, Xinyin Ma, Xinchao Wang

To utilize visual information, Multimodal Large Language Model (MLLM) relies on the perception process of its vision encoder. The completeness and accuracy of visual perception significantly influence the precision of spatial reasoning, fine-grained understanding, and other tasks. However, MLLM still lacks the autonomous capability to control its own visual perception processes, for example, selectively reviewing specific regions of an image or focusing on information related to specific object categories. In this work, we propose the concept of Visual Perception Token, aiming to empower MLLM with a mechanism to control its visual perception processes. We design two types of Visual Perception Tokens, termed the Region Selection Token and the Vision Re-Encoding Token. MLLMs autonomously generate these tokens, just as they generate text, and use them to trigger additional visual perception actions. The Region Selection Token explicitly identifies specific regions in an image that require further perception, while the Vision Re-Encoding Token uses its hidden states as control signals to guide additional visual perception processes. Extensive experiments demonstrate the advantages of these tokens in handling spatial reasoning, improving fine-grained understanding, and other tasks. On average, the introduction of Visual Perception Tokens improves the performance of a 2B model by 23.6\%, increasing its score from 0.572 to 0.708, and even outperforms a 7B parameter model by 13.4\% (from 0.624). Please check out our repo https://github.com/yu-rp/VisualPerceptionToken

📄 PDF Abstract BibTeX arXiv:2502.17425

Code (1)

yu-rp/visualperceptiontoken 공식 구현 pytorch

Tasks

Language ModelingLanguage ModellingLarge Language ModelMultimodal Large Language ModelSpatial Reasoning

Similar Papers 제목 키워드 기반

Spotlight on Token Perception for Multimodal Reinforcement Learning

2025-10-10 · Siyuan Huang, Xiaoye Qu, Yafu Li, Yun Luo 외 arxiv

While Reinforcement Learning with Verifiable Rewards (RLVR) has advanced the reasoning capabilities of Large Vision-Language Models (LVLMs), most existing methods in multimodal reasoning neglect the critical role of visu…

Reinforcement LearningMultimodal Reasoning

Perception Tokens Enhance Visual Reasoning in Multimodal Language Models

2024-12-04 · CVPR 2025 1 · Mahtab Bigverdi, Zelun Luo, Cheng-Yu Hsieh, Ethan Shen 외

Multimodal language models (MLMs) still face challenges in fundamental visual perception tasks where specialized models excel. Tasks requiring reasoning about 3D structures benefit from depth estimation, and reasoning ab…

Depth Estimationobject-detectionObject DetectionVisual Reasoning

Structured Role-Aware Policy Optimization for Multimodal Reasoning

2026-05-08 · Bingqing Jiang, Difan Zou arxiv

Reinforcement learning from verifiable rewards (RLVR), especially with Group Relative Policy Optimization (GRPO), has shown strong potential for improving the reasoning capabilities of large vision-language models (LVLMs…

Reinforcement LearningMultimodal Reasoning

Blink: Dynamic Visual Token Resolution for Enhanced Multimodal Understanding

2025-12-11 · Yuchen Feng, Zhenyu Zhang, Naibin Gu, Yilong Chen 외 arxiv

Multimodal large language models (MLLMs) have achieved remarkable progress on various vision-language tasks, yet their visual perception remains limited. Humans, in comparison, perceive complex scenes efficiently by dyna…

Reasoning Within the Mind: Dynamic Multimodal Interleaving in Latent Space

2025-12-14 · Chengzhi Liu, Yuzhe Yang, Yue Fan, Qingyue Wei 외 arxiv

Recent advancements in Multimodal Large Language Models (MLLMs) have significantly enhanced cross-modal understanding and reasoning by incorporating Chain-of-Thought (CoT) reasoning in the semantic space. Building upon t…

Multimodal ReasoningImage Generation