handwritten digit recognition on the MNIST dataset. Basically, we will perform classification on various handwritten texts and judge whether they are valid digits or not using the MNIST dataset.
But before we see that, we’ll have to learn what MNIST is.
What is MNIST?
Set of 70,000 small images of digits handwritten by high school students and employees of the US causes Bureau. All images are labeled with the respective digit they represent. MNIST is the hello world of machine learning. Every time a data scientist or machine learning engineer makes a new algorithm for classification, they would always first check its performance on the MNIST dataset. There are 70,000 images and each image has 2828 = 784 features. Each image is 2828 pixels and each feature simply represents one-pixel intensity from 0 to 255. If the intensity is 0, it means that the pixel is white and if it is 255, it means it is black.