For this assignment you are to practice image processing, computer vision, and knowledge-creation to try to classify MNIST pictures of handwritten digits 0-9 using Python with Numpy and Scikit Image. Specifically, you are to:

1) Read one image of each digit from the dataset and invert it using only the Numpy library.

2) Use some feature-based approach of your own, or one we will discuss in class, to try to classify the digits. You man *NOT* use any machine-learning technique, we'll discuss those later. Here we are trying to see if we can find a set of features, and to see how far they take us. We are also learning to use Python and the libraries.

3) The MNIST images are used , that have subdirectories, 0,1,2... contain the images for 0's, 1's, 2's respectively. So write the code by defining the path of images.

4) Choose a subset of the provided images to develop your algorithm, but save, say 20% of each to test your final result.