Foundation for Preservation of Knowledge Vedic Library
Dataset Collection:
After placing a cropped image that needs to be added to the dataset, use createVowelDataset.py to generate the inverse binarized and reshaped image which can be added to the collection.
Classfication:
- GeetaDevanagariCustomDataset.py to iterate quickly with custom dataset. It uses completely handbuilt dataset and not use UCI repository at all.
- GeetaDevanagari.py which augments UCI dataset with the custom dataset created. The "path" variable defined in prepareTestTrainData function needs to point to the Train data from UCI dataset