Presentation attack detection

A system for presentation attack detection (PAD) using machine learning, the system comprising:

• a platen for receiving a first object thereon, wherein the platen is transmissive to an optical wavelength of light;
• an illumination module configured to illuminate at least a portion of the first object with illumination from an illumination source, wherein the illumination is spatially patterned prior to reaching the at least a portion of the first object;
• an optical sensing module configured to receive the spatially patterned illumination after the spatially patterned illumination interacts with the at least a portion of the first object;
• and
• a trained PAD machine learning algorithm configured to receive image data based on the spatially patterned illumination received at the optical sensing module and detect whether the image data corresponds to a presentation attack, the PAD machine learning algorithm trained for PAD by:
• obtaining a platen image generated from spatially patterned light of a second object at a platen surface;
• determining at least a subimage of the platen image and dividing the subimage of the platen image into a plurality of patches;
• and
• feeding the plurality of patches to the PAD machine learning algorithm to train the algorithm.