Systems and methods for performing fingerprint based user authentication using imagery captured using mobile devices
Inventors
Mather, Jonathan Francis • Othman, Asem • Tyson, Richard • Simpson, Andrew
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-11188734-B2
Publication Date
2021-11-30
Expiration Date
Abstract
Technologies are presented herein in support of a system and method for performing fingerprint recognition. Embodiments of the present invention concern a system and method for capturing a user's biometric features and generating an identifier characterizing the user's biometric features using a mobile device such as a smartphone. The biometric identifier is generated using imagery captured of a plurality of fingers of a user for the purposes of authenticating/identifying the user according to the captured biometrics and determining the user's liveness. The present disclosure also describes additional techniques for preventing erroneous authentication caused by spoofing. In some examples, the anti-spoofing techniques may include capturing one or more images of a user's fingers and analyzing the captured images for indications of liveness.
Core Innovation
The described invention relates to a mobile-device fingerprint/user authentication system that captures images depicting a finger of a subject using a mobile device having a camera. A plurality of fingers is used for recognition, with finger images acquired and processed to extract discriminatory features and generate biometric identifiers. The system performs authentication/verification and includes liveness detection.
The problem being solved is that fingerprint recognition using a single-finger approach or embedded-sensor approaches can have reduced reliability, including enrollment failures, and that spoof attacks must be resisted. The invention addresses this by using multi-finger evidence fusion and universality against enrollment failures, and by incorporating anti-spoofing measures that require multiple fingers and liveness checks.
In the overall processing architecture, the mobile client captures finger images while a server/cloud system includes modules for capture, analysis, enrollment, authentication, and storing biometric-related data. The processing pipeline includes detecting a portion of a hand that includes a finger using an object detection algorithm, identifying a fingertip segment from the detected portion using a segmentation algorithm, and generating a biometric identifier by extracting discriminatory features from the fingertip segment. The object detection and segmentation are described as being based on classifiers, including HOG, LBP, and Haar feature classifiers.
For liveness detection and anti-spoofing, the document describes liveness checks based on fingerprint ridge reflectivity using flash-on versus flash-off imagery and scoring via high-pass filtering and histogram-based scoring, along with passive quality/color checks and other classifier-based live-vs-spoof approaches. Additional described liveness options include depth-from-focus, depth-from-motion, and gesture-based liveness. Optional extensions are described, including multi-frame capture, normalization and segmentation refinements, combining with other modalities such as face/iris, and additional capture modalities such as NIR/IR.
Claims Coverage
The partial content provided includes two independent claims: a method claim for performing fingerprint recognition and a system claim for performing fingerprint recognition. Across the independent claims, there are five main inventive features: mobile-device finger-image capture, classifier-based object detection of hand and finger regions, classifier-based segmentation to identify a fingertip segment, generation of a biometric identifier from discriminatory features extracted from the fingertip segment, and storage of the biometric identifier in a storage medium.
Mobile-device finger-image capture
Capturing, by a mobile device having a camera, images depicting a finger of a subject.
Classifier-based object detection of hand and finger region
Detecting a portion of a hand that includes at least one finger in an image using an object detection algorithm, wherein the algorithm is a classifier trained to examine regions of an image to detect regions that depict the at least one finger, including HOG, LBP, and Haar feature classifiers.
Classifier-based fingertip segmentation
Identifying a fingertip segment from the image using a segmentation algorithm applied after detecting the portion of the hand, wherein the segmentation algorithm is a classifier trained to detect fingertip segments within the region determined to depict the at least one finger, including HOG, LBP, and Haar feature classifiers.
Biometric identifier from discriminatory fingertip features
Generating a biometric identifier by extracting discriminatory features from the identified fingertip segment.
Storing the generated biometric identifier
Storing the generated biometric identifier in the storage medium.
Both independent claims are directed to mobile-device fingerprint recognition that captures finger images, detects a hand/finger region using a classifier-based object detection algorithm, segments a fingertip segment using a classifier-based segmentation algorithm, generates a biometric identifier by extracting discriminatory features from the fingertip segment, and stores the generated biometric identifier in the storage medium.
Stated Advantages
Improved reliability versus single-finger or embedded-sensor approaches by multi-finger evidence fusion.
Universality against enrollment failures.
Anti-spoof resistance by requiring multiple fingers and performing liveness checks.
Documented Applications
Mobile-device fingerprint/user authentication using captured images and biometric identifiers generated from fingertip discriminatory features, including authentication/verification with liveness detection.
Interested in licensing this patent?