Method and system for activity classification

An activity classifier system, for classifying human activities using two-dimensional (2D) skeleton data comprising joint positions, the system comprising:

• a skeleton preprocessor that transforms the 2D skeleton data into transformed skeleton data by—
• temporally smoothing the joint positions,
• transforming the joint positions to be relative to one of the joint positions,
• scaling the joint positions to a height of a feature of the 2D skeleton data, and
• computing a velocity of each of the joint positions,
• wherein the transformed skeleton data comprising scaled, relative joint positions and joint velocities;
• a gesture classifier comprising a first recurrent neural network that receives the transformed skeleton data, and is trained to identify a most probable gesture of a plurality of gestures;
• and
• an action classifier comprising a second recurrent neural network that receives information from the first recurrent neural network and is trained to identify a most probable action of a plurality of actions,
• wherein the first recurrent neural network is trained on data comprising 2D skeleton sequences with associated gesture labels, and
• wherein the second recurrent neural network is trained with a pre-trained first recurrent neutral network and 2D skeleton sequences with associated action labels.