Classifying Letters with Neural Networks

Sponsored by:

National Action Council for Minorities in Engineering(NACME) Google Applied Machine Learning Intensive (AMLI) at the MORGAN STATE UNIVERSITY

Developed by:

• Alexander Aybar - Morgan State University
• Talaya Sherdon - Morgan State University
• Ishma'il Scott - Morgan State University
• Tyrell Green - Morgan State University

Description

Documents, handwritten letters, fine print, and even in photographs letters are everywhere. Furthermore they’re all in different fonts and sizes. This creates a very challenging task for us to analyze these physical letters and convert them into data. We did this by training a neural network to identify letters through OCR data files. The OCR data files contain letters that have been converted to binary classifications. In order to train this model we used the letters A to J and numbered them from 0 to 9. Then once the model is done we tested the accuracy and precision of the outcome.

Usage instructions

1. Fork this repo
2. Change directories into your project
3. On the command line, type pip3 install requirements.txt
4. ....