Handwritten-Text-Recognition

The model predicts handwritten texts to actual texts. The input to be provided is a picture of an image containing a word and output is the text.

How to create data for use with the model

Suppose given an image:

We can segment the words in the image:

Crop the words into different imaages and feed to the model.

Output in /src/words.txt

Credits: https://github.com/githubharald/SimpleHTR