Character Recognition

This project was an assignment for my data science course I took during my graduate studies at Columbia University. Using a conventional deep neural network, I was able to create a program that recognized handwritten characters. I wrote this program in python.

The current model recognizes just one particular character (“A”). It only outputs whether the input character is an A or not. Character recognition can further be developed to identify fully written sentences. One of the challenges to consider is when the shape of the letter changes when connected to other characters. I can build and train future models that are capable of identifying characters from other languages. Future models can be designed such that they turn analog texts into digital insight. Deep learning can be used to do more than just identify text - rather, it can extract meaning and valuable insight from the text that it is identifying.

The Performance link contains a video demonstration of the code running