arXiv cs.AI· Chen Peng, Shan Huang, Jianbo Huang, Linrui Qin, Qibo Sun, Mengwei Xu·· 6 天前AI 评分29
多模态大模型为何读不出"眯眼文本"：SquintBench 基准与 Squint OPSD 方法
When MLLMs miss what humans read: diagnosing and learning to read squint text
AI 导读
多模态 LLM 会读不出图像粗结构中的文字（即"眯眼文本"）并声称图中无文字，研究者为此构建了 1076 张生产环境审核图像的 SquintBench 基准。
正文 · 原文
Abstract:Multimodal LLMs can fail to read text embedded in an image's coarse structure (text a human recovers by squinting) and report that no text is present. We name this class Squint Text and study it as a recognition deficiency, motivated by cases from a production moderation pipeline. We introduce SquintBench, comprising 1076 production-derived moderation images, and SquintCtrl, a controlled synthetic suite for mechanism diagnosis. Frozen-weight analysis shows that hidden text leaves a coarse signal while glyph identity is weakly represented; identity becomes substantially more accessible when the correct spatial regions are supplied, implicating spatial selection as a major bottleneck. Suppressing high-frequency content also improves reading at a fixed visual-token budget, while visual token count may also contribute. These findings motivate Squint OPSD: a frozen teacher receives a downscaled view and the ground-truth transcription as privileged context, and its token-level distributions supervise a student along on-policy trajectories. At inference, the student reads the original image alone. Squint OPSD reaches 88.61 ANLS on SquintBench, versus 26.50 for the unadapted backbone, 24.44 for the best general-purpose model on the original view, and 68.60 at best for an off-the-shelf model under a downscaled view. These results substantially narrow the gap between available visual evidence and its generated read-out without auxiliary views at inference.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.13465 [cs.CV] |
| (or arXiv:2610.13465v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.13465 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Chen Peng [view email]
[v1]
Fri, 9 Oct 2026 12:02:22 UTC (2,448 KB)
来源：arXiv cs.AI · arxiv.org