Folders CASIA-HWDB_Test and CASIA-HWDB_Train contain pictures of sample data taken from https://www.kaggle.com/pascalbliem/handwritten-chinese-character-hanzi-datasets
Script CSVHandler.py is used for creating test_china.csv and train_china.csv It converts each picture to array, where the values are pixel's represantation in gray scale. Each row is a different .png file of handwritten character. First column "label" is given as a UTF-8 name of chinese letter, which is taken from name of subdirectory, in which character is included.