Image Recognition : a CNN model with MNIST
I have created and trained a convolutional neural network for MNIST.
Starting point
Computer vision is an important branch of machine learning and it is a great place to start learning the core principles of the domain. To create a machine learning model with MNIST (Modified National Institute of Standards and Technology) dataset is, for some sort, the hello world exercise of image recognition. MNIST is a database of 60,000 small square 28x28 pixel grayscale images of handwritten digits and one common goal we can give to the model is to give the hability to recognize/predict any of the ten numerical digits (0-9) starting from an image as input.
You’ll find below this text, my notebook, the model and the demo page of the result of my exercise.