A model was trained using Google handwritten Fonts using a text corpus containing only digits ranging from 0-9. The main aim was to recognize ICR sheets from such trained data. Our model gave an accuracy of 94.6% using Tesseract Version-4.
chandra-cc / tesseract_icr-sheets Goto Github PK
View Code? Open in Web Editor NEWA model was trained using Google handwritten Fonts using a text corpus containing only digits ranging from 0-9. The main aim was to recognize ICR sheets from such trained data. Our model gave an accuracy of 94.6% using Tesseract Version-4.