Document-image-binarization

Using Neural Networks to create do one of the preprocessing steps in OCR i.e Document Image Binarization.

Dataset

The dataset consists of scanned images of telugu novels and their binary images (Binary Images made by using a threshold).

Name		Name	Last commit message	Last commit date
Latest commit History 9 Commits
README.md		README.md
autoencoders.ipynb		autoencoders.ipynb
bin.png		bin.png
conv-final.ipynb		conv-final.ipynb
conv.ipynb		conv.ipynb
dilated-cnn.ipynb		dilated-cnn.ipynb
fd2.ipynb		fd2.ipynb
full_data.ipynb		full_data.ipynb
orig.png		orig.png