US 12,455,897 B2
Information processing apparatus, information processing method, and computer program
Akinori Ebihara, Tokyo (JP); and Taiki Miyagawa, Tokyo (JP)
Assigned to NEC CORPORATION, Tokyo (JP)
Appl. No. 18/269,681
Filed by NEC Corporation, Tokyo (JP)
PCT Filed Dec. 28, 2020, PCT No. PCT/JP2020/049173
§ 371(c)(1), (2) Date Jun. 26, 2023,
PCT Pub. No. WO2022/144992, PCT Pub. Date Jul. 7, 2022.
Prior Publication US 2024/0061859 A1, Feb. 22, 2024
Int. Cl. G06F 16/26 (2019.01)
CPC G06F 16/26 (2019.01) 10 Claims
OG exemplary drawing
 
1. An information processing apparatus comprising:
at least one memory storing instructions; and
at least one processor that is configured to execute the instructions to:
directly obtain time series data of biometric information of a person from a camera or a microphone, the time series data including a plurality of elements, the biometric information being images of a face of the person in a case where the time series data is obtained from the camera, the biometric information being speech of the person in a case in a case where the time series data is obtained from the microphone;
perform a third process of extracting a feature vector from a difference feature vector calculated sequentially as a difference between a first vector extracted from the time series data of a target, wherein the time series data of the target includes at least two consecutive elements in time series, and a second vector extracted from registration data of biometric information of a registered person, to calculate a degree of similarity or difference between the time series data and the registration data;
calculate a likelihood ratio indicating a likelihood that the person of the time series data is a same person as the registered person of the registration data, in consideration of the degree of similarity or difference between the time series data and the registration data, on the basis of the feature vector extracted from the difference feature vector; and
determine whether or not the person of the time series data is the registered person of the registration data, on the basis of the likelihood ratio, to perform biometric authentication of the person, where the biometric authentication is successful when the person has been determined to be the registered person and is unsuccessful when the person has not been determined to be the registered person; and
upon determining that the person of the time series data is the registered person of the registration data, perform a registration process conditioned on the biometric authentication.