CLEAR：释放统一多模态模型中退化图像理解的生成潜力

CLEAR: Unlocking Generative Potential for Degraded Image Understanding in Unified Multimodal Models

April 6, 2026

作者: Xiangzhao Hao, Zefeng Zhang, Zhenyu Zhang, Linhao Yu, Yao Chen, Yiqian Zhang, Haiyun Guo, Shuohuan Wang, Yu Sun

cs.AI

摘要

在现实场景中，因模糊、噪点、压缩和光照不足导致的图像质量退化严重制约着多模态理解能力。将理解与生成功能整合于单一架构的统一多模态模型，因其生成路径能模拟退化所破坏的细粒度视觉结构，天然适合应对这一挑战。然而现有模型未能针对退化输入有效利用其自身生成能力。我们发现这种脱节源于两个叠加因素：现有训练机制从未要求模型在推理过程中调用生成能力，且标准的"解码-再编码"路径无法支持有效的联合优化。本文提出CLEAR框架，通过三个渐进步骤实现两种能力的衔接：(1)在退化感知数据集上进行监督微调，建立"先生成后回答"的推理模式；(2)通过潜在表示桥接技术，用可直接优化的生成-推理直连路径取代解码-再编码的迂回方案；(3)采用交错式GRPO强化学习方法，在答案正确性奖励下联合优化文本推理与视觉生成。我们构建了MMD-Bench评估基准，涵盖六个标准多模态测试集的三种退化严重程度。实验表明CLEAR在保持清晰图像性能的同时，显著提升了对退化输入的鲁棒性。进一步分析发现，移除像素级重建监督后产生的中间视觉状态具有更高感知质量，表明任务驱动优化与视觉质量存在内在一致性。

English

Image degradation from blur, noise, compression, and poor illumination severely undermines multimodal understanding in real-world settings. Unified multimodal models that combine understanding and generation within a single architecture are a natural fit for this challenge, as their generative pathway can model the fine-grained visual structure that degradation destroys. Yet these models fail to leverage their own generative capacity on degraded inputs. We trace this disconnect to two compounding factors: existing training regimes never ask the model to invoke generation during reasoning, and the standard decode-reencode pathway does not support effective joint optimization. We present CLEAR, a framework that connects the two capabilities through three progressive steps: (1) supervised fine-tuning on a degradation-aware dataset to establish the generate-then-answer reasoning pattern; (2) a Latent Representation Bridge that replaces the decode-reencode detour with a direct, optimizable connection between generation and reasoning; (3) Interleaved GRPO, a reinforcement learning method that jointly optimizes text reasoning and visual generation under answer-correctness rewards. We construct MMD-Bench, covering three degradation severity levels across six standard multimodal benchmarks. Experiments show that CLEAR substantially improves robustness on degraded inputs while preserving clean-image performance. Our analysis further reveals that removing pixel-level reconstruction supervision leads to intermediate visual states with higher perceptual quality, suggesting that task-driven optimization and visual quality are naturally aligned.