Recognizing handwritten digits with Machine Learning
A Machine Learning project which recognizes handwritten digits from the MNIST dataset with 97 % accuracy. Written in Python using Numpy, completely from the ground up (no TensorFlow etc). Uses PyQT5 for a visual interface.
Aim
The aim of this project was implement a neural network from scratch to learn the fundamentals of machine learning. It also served as a learning opportunity for the GUI framework QT5.
April 1st, 2020 · William Sandström and Harald Bjurulf