US 12,688,731 B2Granted

Method and system for detection of a fraudulent action using face database search and retrieval

AssigneeHyperverge Inc.

InventorsManideep Kolla, Ayush Pandey, Aravinth Muthu, Hariprasad P S, Vignesh Krishnakumar

Life of this patent

Estimating the term…

Loading patent…

Abstract

A system and method for detection of a fraudulent action using face database search and retrieval is given. The method identifies fraudulent action using a combination of images captured in real-time, feature vector extraction, and metadata matching. The method involves capturing, a query selfie image and a query ID card image. Thereafter, face feature vectors from the query images are extracted, and candidate face feature vectors are identified by searching a set of nearest neighbors of the extracted face feature vectors. The method then retrieves pre-stored application IDs associated with the candidate face feature vectors and a pre-stored metadata associated with the candidate face feature vectors. The method then matches the query metadata with the pre-stored metadata and generates either a fraudulent or non-fraudulent indication based on the matching. Thereafter it detects a fraudulent activity based on the fraudulent indication.

Claims29

Dependent claims

A method for detection of a fraudulent action, the method comprising:

capturing in real time, by a capturing unit, a query selfie image and a query ID card image, wherein the query selfie image comprises a human object, and the query ID card image comprises at least one of the human object and a query metadata;
processing, by a processing unit, the query selfie image and the query ID card image;
extracting, by the processing unit, a query image face feature vector based on the processing of the query selfie image, and a query ID card image face feature vector and the query metadata based on the processing of the query ID card image;
searching, by a searching unit, a set of nearest neighbors of the query image face feature vector in a set of pre-stored selfie face feature vectors, and a set of nearest neighbors of the query ID card image face feature vector in a set of pre-stored ID face feature vectors;
identifying, by the searching unit:
one or more candidate selfie face feature vectors in the set of pre-stored selfie face feature vectors based on the searching of the set of nearest neighbors of the query image face feature vector, and
one or more candidate ID face feature vectors in the set of pre-stored ID face feature vectors based on the searching of the set of nearest neighbors of the query ID card image face feature vector;
retrieving, by the searching unit from a storage unit, a pre-stored selfie application ID associated with the one or more candidate selfie face feature vectors, and a pre-stored ID card application ID associated with the one or more candidate ID face feature vectors;
retrieving, by the searching unit from a metadata database,
a first pre-stored metadata associated with the one or more candidate selfie face feature vectors based on the pre-stored selfie application ID, and
a second pre-stored metadata associated with the one or more candidate ID face feature vectors based on the pre-stored ID card application ID;
matching, by a decision-making unit, the query metadata with at least one of the first pre-stored metadata and the second pre-stored metadata;
generating, by the decision-making unit, one of a fraudulent indication and a non-fraudulent indication based on the matching of the query metadata with at least one of the first pre-stored metadata and the second pre-stored metadata;
and
detecting, by the decision-making unit, the fraudulent action based on the fraudulent indication.

12 more elements

A system for detection of a fraudulent action, the system comprises:

a capturing unit, configured to capture a query selfie image and a query ID card image, wherein the query selfie image comprises a human object, and the query ID card image comprises at least one of the human object and a query metadata;
a processing unit connected at least to the capturing unit, wherein the processing unit is configured to:
process the query selfie image and the query ID card image, and
extract a query image face feature vector based on the processing of the query selfie image, and a query ID card image face feature vector and the query metadata based on the processing of the query ID card image;
a searching unit connected at least to the capturing unit and the processing unit, wherein the searching unit is configured to:
search a set of nearest neighbors of the query image face feature vector in a set of pre-stored selfie face feature vectors, and a set of nearest neighbors of the query ID card image face feature vector in a set of pre-stored ID face feature vectors,
identify one or more candidate selfie face feature vectors in the set of pre-stored selfie face feature vectors based on the searching of the set of nearest neighbors of the query image face feature vector,
identify one or more candidate ID face feature vectors in the set of pre-stored ID face feature vectors based on the searching of the set of nearest neighbors of the query ID card image face feature vector,
retrieve from a storage unit, a pre-stored selfie application ID associated with the one or more candidate selfie face feature vectors, and a pre-stored ID card application ID associated with the one or more candidate ID face feature vectors,
retrieve from a metadata database a first pre-stored metadata associated with the one or more candidate selfie face feature vectors based on the pre-stored selfie application ID, and
retrieve from the metadata database a second pre-stored metadata associated with the one or more candidate ID face feature vectors based on the pre-stored ID card application ID;
and
a decision-making unit connected at least to the capturing unit, the processing unit, and the searching unit, wherein the decision-making unit is configured to:
match the query metadata with at least one of the first pre-stored metadata and the second pre-stored metadata,
generate one of a fraudulent indication and a non-fraudulent indication based on the matching of the query metadata with at least one of the first pre-stored metadata and the second pre-stored metadata, and
detect the fraudulent action based on the fraudulent indication.

13 more elements

A non-transitory computer readable storage medium storing instructions for detection of a fraudulent action, the storage medium comprising executable code which, when executed by a processor, causes the processor to:

enable a capturing unit to capture a query selfie image and a query ID card image, wherein the query selfie image comprises a human object, and the query ID card image comprises at least one of the human object and a query metadata;
process the query selfie image and the query ID card image;
extract a query image face feature vector based on the processing of the query selfie image, and a query ID card image face feature vector and the query metadata based on the processing of the query ID card image;
enable a searching unit to:
search a set of nearest neighbors of the query image face feature vector in a set of pre-stored selfie face feature vectors, and a set of nearest neighbors of the query ID card image face feature vector in a set of pre-stored ID face feature vectors,
identify one or more candidate selfie face feature vectors in the set of pre-stored selfie face feature vectors based on the searching of the set of nearest neighbors of the query image face feature vector,
identify one or more candidate ID face feature vectors in the set of pre-stored ID face feature vectors based on the searching of the set of nearest neighbors of the query ID card image face feature vector,
retrieve from a storage unit, a pre-stored selfie application ID associated with the one or more candidate selfie face feature vectors, and a pre-stored ID card application ID associated with the one or more candidate ID face feature vectors,
retrieve from a metadata database a first pre-stored metadata associated with the one or more candidate selfie face feature vectors based on the pre-stored selfie application ID, and
retrieve from the metadata database a second pre-stored metadata associated with the one or more candidate ID face feature vectors based on the pre-stored ID card application ID;
and
enable a decision making unit to:
match the query metadata with at least one of the first pre-stored metadata and the second pre-stored metadata,
generate one of a fraudulent indication and a non-fraudulent indication based on the matching of the query metadata with at least one of the first pre-stored metadata and the second pre-stored metadata, and
detect the fraudulent action based on the fraudulent indication.

12 more elements

Dependencies are inferred from the claim text.

Facts

Application: 18/336,504Filed 16 Jun 2023 · us provisional application
Priority: 5 Dec 2022
Published: US 2024/0185635 A16 Jun 2024
Granted: 21 Jul 2026
Term ends: Not estimatedThe source records don’t support an estimate.

Classification8 codes

G06V 40/179
Human faces, e.g. facial parts, sketches or expressions: metadata assisted face recognition
G06V 10/24
Image preprocessing: aligning, centring, orientation detection or correction of the image
G06V 10/25
Image preprocessing: determination of region of interest [ROI] or a volume of interest [VOI]
G06V 10/82
Using pattern recognition or machine learning: using neural networks