File list:

annotations.zip - contains raw annotation data in JSON format (annotations.json).

img_urls.txt - list of URLs of images in the Naxi manuscripts collection corresponding to the annotations. Use this to obtain the images.

keywords_B-C_2023.9.29.txt - list of keywords assigned to the pictograms in the dictionaries by Fang (1979) and Li (2001), excluding those copied from the dictionaries.

ml_experiments.zip - contains an interactive Python notebook (dongba_char_recognition.ipynb) as well as the training and test data in COCO format which can be used to reproduce the machine learning experiment described in the thesis.