任务效用IQA：重新思考具身任务中的图像质量评估
原标题：Task-Utility IQA: Rethinking Image Quality Assessment in Embodied Tasks
AI 摘要
该论文提出将图像质量评估（IQA）从任务无关的分数回归转向以具身任务效用为目标的条件化判断，并给出基于多模态大语言模型（MLLM）的具身IQA智能体框架。作者通过VisDrone证据研究指出，传统感知指标和通用MLLM的IQA分数难以预测下游视觉语言模型任务的成功，而任务条件化的效用分数预测效果更好。论文主张IQA应面向具身、可解释和决策感知的视觉感知重新定义，并承诺开源代码与实验结果。
正文节选
tcboxmath \tl_set:Ne\tcbhighmathtcbhighmath Task-Utility IQA: Rethinking Image Quality Assessment in Embodied TasksThanks: These authors contributed equally. † Corresponding author. Abstract Image quality assessment (IQA) underpins embodied imaging pipelines by judging whether visual quality satisfies downstream tasks, yet most existing methods learn task-agnostic scores aligned with generic human ratings on static benchmarks. This objective mismatches embodied and interactive settings, where im