Multi-View Face Recognition for Payment Security
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current online payment systems face challenges in accurately verifying user identities, which compromises the security of electronic accounts, as they rely primarily on frontal face images for recognition.
Innovation Solution
The method involves acquiring both frontal and side face images of a user, combining them for identity recognition, using key point detection models and convolutional neural networks to enhance feature extraction and similarity comparison, thereby improving recognition accuracy and security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If only frontal face image is used for identity recognition, then the operation is simple and fast, but the recognition accuracy and security are insufficient
Solution Approach 1:
The patent transitions from using only frontal face images (2D plane recognition) to incorporating side face images (adding depth dimension), thereby moving from single-view to multi-view recognition. This dimensional expansion provides more comprehensive facial features including profile contours, ear shapes, and nasal structures, significantly improving identity recognition accuracy and security while maintaining operational simplicity through automated multi-angle image capture.
2Reliability
If multiple face images are collected and processed, then more feature information is obtained for better security, but the processing time and computational complexity increase
Solution Approach 1:
The patent implements preliminary action by pre-collecting and storing multiple face images (frontal and side views) during account registration. These pre-acquired images are processed in advance to extract key facial features and create recognition templates. During actual payment verification, the system only needs to compare newly captured images against these pre-processed templates, significantly reducing real-time processing time while maintaining high security standards through multi-angle feature matching.
Data Source
AI summary
A payment method includes: acquiring a frontal face image, and a side face image from at least one side, of a to-be-identified user; identifying whether the to-be-identified user is an authorized user of a payment electronic account according to the frontal face image and the side face image that are acquired; and making a deduction operation on the payment electronic account if it is identified that the to-be-identified user is an authorized user of the payment electronic account.