TactileEval: 向自动细粒度评估和编辑触觉图形迈出一步
TactileEval: A Step Towards Automated Fine-Grained Evaluation and Editing of Tactile Graphics
AI总结：
本文提出TactileEval，通过三阶段流程实现触觉图形的自动细粒度评估与编辑，基于专家评论建立质量分类体系，并利用ViT模型实现高精度任务识别与针对性修正。
摘要：
触觉图形在到达盲人和视障学习者前需要仔细专家验证，但现有数据集仅提供粗粒度的整体质量评分，无法提供可操作的修复信号。本文提出TactileEval，一个三阶段流程，旨在自动化这一过程。基于TactileNet数据集的专家自由文本评论，我们建立了五类质量分类体系，包括视角、部分完整性、背景杂乱、纹理分离和线质量，符合BANA标准。随后通过Amazon Mechanical Turk收集了14,095个结构化注释，涵盖66个对象类别，分为六个不同类别。一个可重复使用的ViT-L/14特征探针在该数据上训练，实现了30种不同任务的85.70%总体测试准确率，一致的难度排序表明该分类体系捕捉到了有意义的感知结构。基于这些评估，我们提出了一个基于ViT的自动编辑流程，将分类器评分通过类别特定的提示模板路由，通过gpt-image-1图像编辑生成针对性修正。代码、数据和模型可在https://TactileEval.github.io/获取。
英文摘要：
Tactile graphics require careful expert validation before reaching blind and visually impaired (BVI) learners, yet existing datasets provide only coarse holistic quality ratings that offer no actionable repair signal. We present TactileEval, a three-stage pipeline that takes a first step toward automating this process. Drawing on expert free-text comments from the TactileNet dataset, we establish a five-category quality taxonomy; encompassing view angle, part completeness, background clutter, texture separation, and line quality aligned with BANA standards. We subsequently gathered 14,095 structured annotations via Amazon Mechanical Turk, spanning 66 object classes organized into six distinct families. A reproducible ViT-L/14 feature probe trained on this data achieves 85.70% overall test accuracy across 30 different tasks, with consistent difficulty ordering suggesting the taxonomy suggesting the taxonomy captures meaningful perceptual structure. Building on these evaluations, we present a ViT-guided automated editing pipeline that routes classifier scores through family-specific prompt templates to produce targeted corrections via gpt-image-1 image editing. Code, data, and models are available at https://TactileEval.github.io/
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
