Belfort

该数据集是一个专为手写文本识别而创建的行级别数据集，其来源于法国贝尔福市市政登记册。它包含了每行文本的多个转录版本。此外，该数据集还包含了来自多位人类标注者以及自动手写识别模型的标注信息。该数据集的规模为24,105行文本，其任务是对手写文本进行识别。

This is a line-level dataset specifically developed for handwritten text recognition, sourced from the municipal registers of Belfort, France. It includes multiple transcript versions for each line of text. Additionally, the dataset contains annotation information from multiple human annotators as well as automatic handwritten recognition models. The dataset comprises 24,105 lines of text, with the core task being handwritten text recognition.