Back to results
US 12,764,108 B2Granted
Multi-factor identity verification
AssigneeMyHealthID, Inc.
InventorsJay Nitturkar, Erika Eördögh
Life of this patent
Estimating the term…
Abstract
Provided are embodiments of systems, devices and methods for multi-factor identity verification, which may include utilization of automated picture ID to Selfie matching, cross-reference address information, biometrics and geo-location information and unique smartphone device identifiers, especially in the context of healthcare industry.
Claims20
A computer-implemented method for verifying an identity of a patient, the method comprising:
- receiving, from a client device, insurance card data extracted by optical character recognition, the insurance card data including at least a patient name, an address, and a date of birth;
- receiving, from the client device, driver's license data extracted by optical character recognition, the driver's license data including at least a license number, a name, an address, a date of birth, and an extracted headshot image;
- receiving, from the client device, a selfie image of a user;
- determining a face-match score between the selfie image and the extracted headshot image;
- determining an address-match score based on comparing the insurance card data to the driver's license data;
- determining a date-of-birth concordance class based on comparing the date of birth from the insurance card data to the date of birth from the driver's license data;
- receiving one or more contextual signals from the client device, the contextual signals including at least one of a device geolocation, a timestamp, a timezone, a device fingerprint, or an appointment-window indicator;
- generating an identity evidence vector comprising the face-match score, the address-match score, the date-of-birth concordance class, and the contextual signals;
- computing, using a scoring function, a risk score based on the identity evidence vector;
- applying, to the risk score, a verification policy specifying a k-of-N rule requiring satisfaction of at least k identity conditions among N evaluated identity conditions;
- in response to determining that the verification policy is satisfied, generating a verification token containing at least indicators of address agreement, facial match, and date-of-birth agreement;
- in response to determining that the verification policy is not satisfied, initiating an escalation workflow including generation of one or more audit artifacts and reason codes.
9 more elements
A system for multi-factor patient identity verification, comprising:
- one or more processors;
- and
- memory storing instructions that, when executed by the one or more processors, cause the system to:
- receive insurance-card OCR data including a name, an address, and a date of birth;
- receive driver's-license OCR data including a license number, a name, an address, a date of birth, and an extracted headshot;
- receive a selfie image;
- generate an identity evidence vector including a face-match score between the selfie and the extracted headshot, an address-match score, a date-of-birth concordance class, and one or more contextual signals;
- compute a risk score based on the identity evidence vector;
- apply a verification policy including a k-of-N rule;
- generate a verification token when the policy is satisfied;
- and
- generate an audit trail with reason codes when the policy is not satisfied.
9 more elements
A non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause a device to perform operations comprising:
- receiving OCR-extracted insurance card data including a name, an address, and a date of birth;
- receiving OCR-extracted driver's license data including a name, an address, a date of birth, and an extracted headshot;
- receiving a selfie image;
- computing a face-match score between the selfie and the extracted headshot;
- computing an address-match score and a date-of-birth concordance class;
- calculating a risk score using a scoring function applied to the face-match score, the address-match score, the date-of-birth concordance class, and contextual signals;
- determining whether the risk score satisfies a k-of-N verification policy;
- generating a verification token when the policy is satisfied;
- and
- generating an audit trail when the policy is not satisfied.
7 more elements
Dependencies are inferred from the claim text.