Manually validated PageXML files for images in book "Lettres du sieur de Balzac"

Transcription of 57 pages of the 637-page printed book "Lettres du sieur de Balzac" by Jean-Louis de Balzac (1597-1654), publication date 1624. Transcription contains pages in PageXML format, useful for training an optical character recognition (OCR) model. The PageXML files were created by applying a public multilingual Transkribus model ("Transkribus print M1") on the images at https://github.com/Heresta/OCR17plus/blob/main/Data/Balzac1624_Lettres_btv1b86262420_corrected and by manually validating the result. During manual validation, the transcription was normalized (spelling adapted to modern French).本数据集转录了让-路易·德·巴尔扎克（Jean-Louis de Balzac，1597-1654）所著637页印刷版典籍《Lettres du sieur de Balzac》（《巴尔扎克先生书信集》）中的57页，该书刊行于1624年。该转录数据采用PageXML格式存储，可用于训练光学字符识别（Optical Character Recognition，OCR）模型。上述PageXML文件通过将公开的多语言Transkribus模型（"Transkribus print M1"）应用于指定图像集（https://github.com/Heresta/OCR17plus/blob/main/Data/Balzac1624_Lettres_btv1b86262420_corrected）生成，并经人工校验输出结果完成最终校准。在人工校验阶段，转录文本已完成归一化处理（拼写调整为现代法语规范）。