Character Recognition Models

• This project implements a neural network from scratch in Python without the use of higher-level frameworks like Pytorch and TensorFlow.
• The only libraries used are Numpy, OpenCV, Pickle, Matplotlib and Flask.
• Accuracy obtained was 95% with this English Handwritten Characters Dataset: Kaggle Dataset

References

• [1] “MNIST handwritten digit database, Yann LeCun, Corinna Cortes and Chris Burges,” yann.lecun.com. https://yann.lecun.com/exdb/mnist/
• [2] Cohen, G., Afshar, S., Tapson, J., & van Schaik, A. (2017). EMNIST: an extension of MNIST to handwritten letters.