Used CNN model on the MNIST dataset to detect the digit with 97% accuracy. To use the model, create an image on paint and save it to the same directory, and enter the name of the image to run the model. To run the file, download the mnist_train dataset from: https://www.kaggle.com/oddrationale/mnist-in-csv.
Used CNN model and trained it on a portion of the A-Z Handwritten Alphabet dataset on Kaggle. Trained to have around 74% accuracy. Alternative: Used pytessersact and OCR to detect handwritten text.