Handwritten Digit Recognition - Machine Learning Project

Project Overview

I wanted a model that could recognize handwritten digits beyond the MNIST examples. I trained it with TensorFlow and Keras, processed images with NumPy and OpenCV, and tested it on digits I drew in Paint. The model recognized handwritten input through the program's visual interface.

Key Features

• MNIST training data
• TensorFlow and Keras model
• OpenCV image input

What I Learned

I learned how NumPy represents image pixels, how OpenCV supplies input, and how TensorFlow's Keras API trains the classifier. Testing my own drawn digits helped me see why consistent image preparation matters beyond the MNIST dataset.