Systems and methods for human pose and shape recovery

A system comprising one or more processors and one or more storage devices storing instructions that, when executed by the one or more processors, cause the one or more processors to:

• implement an artificial neural network;
• provide information regarding one or more joint locations of a person to the artificial neural network;
• determine, via the artificial neural network, a first plurality of parameters associated with a pose of the person based on the information regarding the one or more joint locations of the person;
• determine, via the artificial neural network, a second plurality of parameters associated with a shape of the person based on the information regarding the one or more joint locations of the person;
• and
• generate, based on the first and second pluralities of parameters, a human model that represents the pose and shape of the person;
• wherein the artificial neural network is trained to determine the first and second pluralities of parameters and the training of the artificial neural network comprises:
• the artificial neural network receiving training data that comprise joint locations of a human body;
• the artificial neural network predicting pose and shape parameters associated with the human body based on the joint locations comprised in the training data;
• the artificial neural network inferring joint locations of the human body based on the predicted pose and shape parameters;
• and
• the artificial neural network adjusting one or more of its operating parameters based on a difference between the inferred joint locations and the joint locations comprised in the training data.