Extracts motion signals from a video for an object using code based on The Visual Microphone and Phase-based Video Motion Processing. A video data set is used to train a gaussian SVM classifier and a different testing set is used to test the model. The model achieved an accuracy of 70.8% against the testing set.