Handwriting Recognition Project
Recognizes and Digitizes different types of handwriting using Convolutional Recurrent Neural Network (CRNN)
Created on 24th September 2022
•
Handwriting Recognition Project
Recognizes and Digitizes different types of handwriting using Convolutional Recurrent Neural Network (CRNN)
The problem Handwriting Recognition Project solves
Character Recognition utilizes image processing technologies to convert characters on scanned documents into digital forms. It typically performs well in machine-printed fonts. However, it still poses difficult challenges for machines to recognize handwritten characters, because of the huge variation in individual writing styles.
The goal is to create a deep learning project which will recognize the handwritings and detect them with an accuracy of more than 85%.
Challenges I ran into
- Choosing the perfect model was the hardest part as it was time consuming to make a model and test for it's accuracy.
- Had to learn new topics like how to introduce non-linearity in the convolutional layers. Also studied about CTC Loss.
- Had to deal with errors while setting up ML environment on my system.
Technologies used
Discussion
Builders also viewed
See more projects on Devfolio