Face Recognition With Heterogeneous Graphs for Similar-Face Payments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Face recognition technologies face challenges in accurately distinguishing between similar faces, particularly in scenarios like face scan payments, leading to reduced accuracy and user experience.
Innovation Solution
A face recognition method that utilizes a heterogeneous network graph to combine face image features with device and object graph features, calculating an initial probability and generating a fused probability for resource transfer verification, enhancing accuracy and convenience by leveraging historical transfer data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional face recognition is used to distinguish similar faces, then the system is simple to operate, but the recognition accuracy deteriorates
Solution Approach 1:
The patent merges face recognition with heterogeneous graph analysis by combining face image features, device features, and user behavior features into a unified verification system. The graph neural network integrates multiple data sources (face features, device identification, historical transfer records) to jointly determine verification methods, resolving the contradiction by combining multiple simple features into a complex analytical model that achieves high accuracy without requiring complex individual components
Solution Approach 2:
The patent transitions from traditional 2D face image analysis to multi-dimensional verification by incorporating device features, user behavior patterns, and historical transfer data as additional dimensions. The heterogeneous graph structure adds semantic dimensions (device-object relationships, user-device associations) beyond visual face features, enabling accurate distinction of similar faces through multi-dimensional feature space expansion
2Measurement precision
If multi-dimensional verification is implemented to improve accuracy, then recognition precision improves, but operation convenience deteriorates
Solution Approach 1:
The patent implements dynamic verification method selection that adapts to each specific scenario. The system dynamically determines the appropriate verification level (from simple face recognition to enhanced verification) based on real-time analysis of face similarity, device characteristics, and user behavior patterns. This dynamic adaptation ensures high accuracy when needed while maintaining convenience for routine transactions, resolving the contradiction between precision and ease of operation
Solution Approach 2:
The patent applies different verification strengths locally based on risk assessment. For high-risk scenarios (similar faces, new devices), enhanced verification with multiple features is applied. For low-risk scenarios (verified users, recognized devices), simple face recognition suffices. This localized quality adjustment ensures maximum accuracy where needed while preserving convenience elsewhere, balancing the contradiction between verification precision and operational ease
Data Source
AI summary
Embodiments of this application disclose a face recognition method and apparatus, an electronic device and a storage medium. The embodiments of this application can obtain a face image feature of a resource transfer object and a device identification of a resource transfer device; search for a target face feature matching the face image feature, and determine a target object corresponding to the target face feature; search for a resource transfer device graph feature corresponding to the device identification and a target object graph feature corresponding to the target object; determine an initial resource transfer probability that the target object performs resource transfer at the resource transfer place; generate a fused resource transfer probability according to a similarity between the face image feature and the target face feature and the initial resource transfer probability; and determine a resource transfer verification manner of the resource transfer object.