This project's objective is to allow ocr of any language and font (handwriting also). The basic idea is instead of relaying on a general algorithm trained on huge datasets, in this project the training (or transfer learning) will be done on the current dataset (i.e. first pages of a book)
- a basic tool for marking classifying and viewing the data is ready.
- a basic NN model for detecting letters added (based on EAST word detection network)
- a basic NN model for identifying letters added (simplest vanilla cnn used)
- for gui tkinter was used, for NN pytorch was used
- improve mvc
- add support for moving letters
- support marking just part of page
- add visualization for training and inference process
- add duplication detection
- investigate strange loss graphs
- chose wisely networks
- add gt page visualization (boxes as image)
- ignore misses and false of lettres detection in boundary
- split letters by logic in train
- add automatic letters clustering
- add support for punctuation (also nikud like in hebrew)
- add lettres on page into words and lines tool
- detect spaces