Machine learning using RNN LSTM with Keras on UTD-MHAD Kinect dataset (Depth, Inertial, Skeleton). This is an assignment from NUS ISS SenseMaking course.

Please refer to jupyter notebook file "activity_classification.ipynb", here we choose only inertial dataset and 3 actvities (hand swipe, two hands push, jogging) for sequence classification.

Run command python depth_data_visualizer.py in terminal launch visualizer tool. User can select action, subject and trial to show the 240 x 320 depth videos