The object of this project is to develop character classifiers for several historical documents, each produced shortly after the advent of the printing press. Documents are in different languages, use different fonts, and have other idiosyncrasies specific to their author and printer. The task is to identify for a given bitmap image of a character the identity of that character.
KNN, Random Forest, CNN are applied to this task. The previous two use python sklearn package, the last one is by Tensorflow.