多域低资源光学字符识别的多专家路由：以满文为例
Multi-Expert Routing for Multi-Domain Low-Resource OCR: A Manchu Case Study
浏览论文内容
中文总结 AI 辅助
针对满文OCR中书写风格多样且标注数据有限的问题，提出多专家系统，复用微调检查点为专家，用图像分类器分配页面，缺专家时额外训练，在测试集上有高精度匹配，报告相关内容使比较可重复。
中文摘要 AI 辅助
历史满文光学字符识别（OCR）必须处理各种视觉上不同的书写风格，包括楷书、行书和用于宫廷奏折的半草书，尽管标注数据有限。我们研究了一种多专家系统，该系统将迭代微调过程中的检查点作为领域专家重复使用，并使用轻量级页面级图像分类器按视觉风格分配页面。当检查点池缺少合适的专家时，我们为该领域训练额外的专家。在三个冻结测试集上，路由系统在两位小数精度上与每种风格的选定专家匹配：楷书的字符错误率（CER）为0.30%，奏折为1.57%，行书为4.83%。路由器实现了99.3%的页面级领域准确率，并以相同精度与领域标签预言机匹配。三个选定专家中的两个并非专门针对其最终领域进行训练；只有行书专家是以该领域为目标进行训练的。我们报告了评估协议、路由器设计和每页预测，以使比较具有可重复性。
英文摘要
Historical Manchu OCR must accommodate various visually distinct writing styles, including regular script, running script, and the semi-cursive chancery hand used in palace memorials, despite limited labeled data. We study a multi-expert system that reuses checkpoints from an iterative fine-tuning process as domain specialists and uses a lightweight page-level image classifier to dispatch pages by visual style. When the checkpoint pool lacks a suitable specialist, we train an additional expert for that domain. On three frozen test sets, the routed system matches the selected specialist for each style at two-decimal precision: 0.30 percent CER on regular script, 1.57 percent on memorials, and 4.83 percent on running script. The router achieves 99.3 percent page-level domain accuracy and matches the domain-label oracle at the same precision. Two of the three selected specialists were not trained specifically for their final domain; only the running-script expert was trained with that domain as its target. We report the evaluation protocol, router design, and per-page predictions to make the comparison reproducible.
发表机构
- Institute of Advanced Studies, Beijing Normal University（北京师范大学高等研究院）
- Faculty of Humanities and Social Sciences, Beijing Normal University–Hong Kong Baptist University United International College (UIC)（北京师范大学-香港浸会大学联合国际学院人文与社会科学学部）
- Department of Applied History, National Chiayi University（国立嘉义大学应用历史系）
机构由 AI 辅助整理，请以论文原文为准。
