arXiv cs.AI· Junjie Xu, Yuzhe Yang, Rui Yang, Luwei Xiao, Wenxin Wu, Xingjiao Wu, Liang He·· 12 小时前AI 评分31
HealBench：评估文生图生成心理舒适度的细粒度基准
HealBench: A Fine-Grained Benchmark for Evaluating Psychological Comfort in Text-to-Image Generation
AI 导读
研究者提出 HealBench，一个包含 30,000 个高质量图文对、覆盖 19 个类别和 92 个子类别的细粒度基准，用于评估文生图输出的心理舒适度。基于 1,095 个实例的 HealBench-Eval 对 14 个代表性图像生成模型和提示词改写策略的评测显示，语义对齐与美学质量并不能保证图像带来情感疗愈效果。
正文 · 原文
Abstract:Despite advances in photorealism, semantic fidelity, and aesthetics, text-to-image generation and image editing remain understudied in a human-centered dimension: whether their outputs feel psychologically comforting. Grounded in cognitive evaluation theory, we define psychological comfort as an affective response shaped by viewers' appraisal of restorative and emotionally supportive visual cues. We introduce HealBench, a fine-grained benchmark of 30,000 high-quality image-text pairs spanning 19 categories and 92 subcategories, selected from 100,000 candidates through a human- and expert-in-the-loop pipeline combining vision-language model matching, human preference models, expert filtering, and exemplar-guided augmentation. We also construct HealBench-Eval, a feature-clustered evaluation set of 1,095 instances. Using HealBench-Eval, we conduct objective and subjective evaluations of 14 representative image generation models and prompt-rewriting strategies. Results reveal a gap between conventional visual quality and perceived comfort: strong semantic alignment and aesthetics do not guarantee emotionally restorative images. Reinforcement learning experiments further show that HealScorer can serve as a reward signal for assessing and improving the emotional and restorative quality of generated images. These findings establish psychological comfort as a distinct dimension of human-centered image generation and position HealBench as a foundation for affect-aware generative models.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.27516 [cs.CV] |
| (or arXiv:2610.27516v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.27516 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Junjie Xu [view email]
[v1]
Tue, 20 Oct 2026 02:31:31 UTC (35,004 KB)
来源：arXiv cs.AI · arxiv.org