Employs neural networks to classify images into four categories: ship, automobile, dog or frog

Viterbi_1.py uses a classic shallow classic network with an accuracy of ~62%.

Viterbi_2.py uses a modern network with L2 Regulation and Convolutional Neural Networks with an accuracy of ~ 83%.

Viterbi_3.py is adapted to the testing data set using a specific combination of convolutional neural networks and activation functions with an accuracy of ~87%. (This could be improved upon with further testing).