Manchu OCR and Translation AI

The Manchu OCR project aims to build a reliable, high-accuracy system capable of automatically reading and segmenting Manchu script from real historical sources. Building on a custom word-segmentation workflow and a large corpus of synthetic and real-world scanned materials collected from multiple repositories, the project has already produced a fine-tuned vision-language model that performs strongly on real-world scanned data. Early tests demonstrate that the system can accurately recognize printed and well-written archival pages, validating the feasibility of applying modern OCR techniques to a low-resource script with complex vertical structures and variable calligraphic styles.