System and method for biometric authentication in connection with camera-equipped devices

An anti-spoofing system for providing selective access to resources available in connection with a smartphone based on a deep learning method, the system comprising:

• (a) the smartphone comprising a digital processor, a memory module, an operating system, and non-transitory storage media comprising instructions executable by the digital processor;
• (b) at least one camera associated with the smartphone and configured to capture a plurality of images comprising a flash image and a non-flash image of a human anatomical feature of a first user, wherein the plurality of images are captured without the human anatomical feature physically contacting the smartphone;
• and
• (c) at least one flash component associated with the smartphone and configured to emit a flash during the capture of the flash image and to not emit a flash during the capture of the non-flash image;
• wherein the digital processor performs the following:
• (1) using sliding-window-based local classifiers and classifiers formed by discriminative local features learned from a collection of exemplar images to analyze the flash image and the non-flash image of the human anatomical feature to label anatomical and non-anatomical pixels by classification scores;
• (2) using a trained detector to locate and segment a region of interest of the human anatomical feature within the flash image and the non-flash image;
• (3) using small sampling windows within the region of interest to identify image patches encompassing characteristic biometric features of the human anatomical feature;
• (4) extracting a signature from pixel-level data of the image patches, wherein the signature is unique to the first user and is used to distinguish the first user from a second user, and wherein extracting the signature comprises creating a histogram of edge orientations in a plurality of the image patches;
• (5) determining a spoofing attempt based on a three dimensional characteristic of the human anatomical feature and a plurality of stored user models, wherein the three dimensional characteristic is based on a comparison between the flash image and the non-flash image, wherein the comparison between the flash image and the non-flash image comprises determining a highlighted region and a shadow region in at least one of the flash image and the non-flash image, to determine whether at least one of the flash image and the non-flash image represents a 3-D object or a 2-D representation of a human anatomical feature;
• (6) denying access to one or more of the resources to the first user, based upon the determination of the spoofing attempt;
• and
• (7) storing the at least one of the plurality of images of the human anatomical feature and the three dimensional characteristic in the collection of exemplar images.

An anti-spoofing method implemented by a smartphone for providing selective access to resources available in connection with a smartphone based on a deep learning method, the method comprising:

• (a) using at least one camera associated with the smartphone and configured to capture a plurality of images comprising a flash image and a non-flash image of a human anatomical feature of a first user, wherein the plurality of images are captured without the human anatomical feature physically contacting the smartphone;
• (b) using at least one flash associated with the smartphone and configured to emit a light during the capture of the flash image;
• (c) using sliding-window-based local classifiers and classifiers formed by discriminative local features learned from a collection of exemplar images to analyze the flash image and the non-flash image of the human anatomical feature to label anatomical and non-anatomical pixels by classification scores;
• (d) using a trained detector to locate and segment a region of interest of the human anatomical feature within the flash image and the non-flash image;
• (e) using small sampling windows within the region of interest to identify image patches encompassing characteristic biometric features of the human anatomical feature;
• (f) extracting a signature from pixel-level data of the image patches, wherein the signature is unique to the first user and is used to distinguish the first user from a second user, and wherein extracting the signature comprises creating a histogram of edge orientations in a plurality of image patches;
• (g) determining a spoofing attempt based on a three dimensional characteristic of the human anatomical feature and a plurality of stored user models, wherein the three dimensional characteristic is based on a comparison between the flash image and the non-flash image, wherein the comparison between the flash image and the non-flash image comprises determining a highlighted region and a shadow region in at least one of the flash image and the non-flash image, to determine whether at least one of the flash image and the non-flash image represents a 3-D object or a 2-D representation of a human anatomical feature;
• (h) denying access of one or more of the resources to the first user based upon the determination of the spoofing attempt;
• and
• (i) storing the at least one image of the human anatomical feature in the collection of exemplar images.