I introduced a method to classify videos of different actions using motion history images (MHI) and their Hu moments. The Hu moment values are used as features to train and test two different supervised learning algorithms, K-nearest neighbors (KNN), and Adaboost.