US 12,033,422 B1
Ultrasonic fingerprint anti-spoofing based on A-line data
Adi Hendel, Giv'atayim (IL); Nathan Altman, San Diego, CA (US); Bence Major, Amsterdam (NL); Javier Frydman, Tel-Mond (IL); and Hasib Siddiqui, San Diego, CA (US)
Assigned to QUALCOMM Incorporated, San Diego, CA (US)
Filed by QUALCOMM Incorporated, San Diego, CA (US)
Filed on Mar. 20, 2023, as Appl. No. 18/186,865.
Int. Cl. G06K 9/00 (2022.01); G06F 21/32 (2013.01); G06V 10/82 (2022.01); G06V 40/12 (2022.01); G06V 40/13 (2022.01); G06V 40/40 (2022.01); G06V 40/50 (2022.01)
CPC G06V 40/1306 (2022.01) [G06F 21/32 (2013.01); G06V 10/82 (2022.01); G06V 40/1376 (2022.01); G06V 40/45 (2022.01); G06V 40/50 (2022.01)] 20 Claims
OG exemplary drawing
 
11. A method, comprising:
controlling, by a control system, an ultrasonic fingerprint sensor system to obtain current A-line data from a target object on an outside surface of an apparatus that includes the ultrasonic fingerprint sensor system, the current A-line data comprising fingerprint sensor data corresponding to reflections of ultrasonic waves from the target object detected by a single receiver pixel or by a cluster of adjacent receiver pixels;
controlling, by the control system, the ultrasonic fingerprint sensor system to obtain current ultrasonic fingerprint image data corresponding to reflections of ultrasonic waves from a target object surface;
obtaining, by the control system and from a memory system, previously-obtained A-line data, the previously-obtained A-line data having previously been obtained by the ultrasonic fingerprint sensor system and from an authorized user;
obtaining, by the control system and from the memory system, previously-obtained ultrasonic fingerprint image data, the previously-obtained ultrasonic fingerprint image data having previously been obtained by the ultrasonic fingerprint sensor system and from the authorized user; and
estimating, by the control system and based at least in part on the current A-line data, the previously-obtained A-line data, the current ultrasonic fingerprint image data and the previously-obtained ultrasonic fingerprint image data, whether the target object is a finger of the authorized user, wherein estimating whether the target object is a finger of the authorized user involves an anti-spoofing process performed by one or more neural networks including a deep neural network (DNN) and a plurality of multilayer perceptrons (MLPs), each of the MLPs being configured to provide output features to the DNN corresponding to a frame of current A-line data and a frame of previously-obtained A-line data.