Simple neural net for handwritten digits recognition
This is C++ implementation of simple neural net with single hidden layer and stochastic gradient descent learning algorithm. This project uses OpenCV for matrix manipulation. There is no good reason to use OpenCV instead of any other library, but I decided to do so in order to learn OpenCV framework.
I train and evaluate net based on MNIST dataset (taken from this Kaggle competition https://www.kaggle.com/c/digit-recognizer)