Face Recognition Using Heterogeneous Graphs for Similar Faces
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 and apparatus that integrates feature extraction using a heterogeneous graph and object graph feature databases and a heterogeneous network graph, utilizing a heterogeneous network graph to combine face image features with device and object graph features to determine an initial transfer probability, and then fusing this with similarity measures to enhance verification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional face recognition is used, then the system is simple to operate, but the recognition accuracy decreases for similar faces
Solution Approach 1:
The patent combines multiple feature types (face image features, device graph features, object graph features) into a unified heterogeneous graph structure. This merging of different data sources and feature representations enables more accurate distinction between similar faces by capturing relationships across multiple dimensions, thereby resolving the contradiction between recognition accuracy and system complexity.
Solution Approach 2:
The patent creates a composite feature representation by integrating face image features with graph features from device and object relationships. This composite approach类似于composite materials, combines different types of information (visual features + relational features) to create a more robust and accurate recognition system that can distinguish similar faces better.
2Measurement precision
If heterogeneous graph integration is used, then face recognition accuracy improves, but computational complexity increases
Solution Approach 1:
The patent performs preliminary feature extraction and graph construction before the actual recognition process. By pre-processing and organizing data into heterogeneous graph structures in advance, the system reduces the computational burden during real-time recognition, thus balancing accuracy improvement with computational efficiency.
Solution Approach 2:
The patent replaces traditional mechanical face matching algorithms with a graph-based computational approach. By substituting direct image comparison with graph feature integration and similarity calculations, the system achieves higher accuracy while managing computational complexity through more efficient information representation and processing.
3Measurement precision
If graph feature integration is used, then verification accuracy increases, but processing time increases
Solution Approach 1:
The patent extracts graph features and constructs the heterogeneous graph structure in advance, before the verification process. This preliminary preparation of feature data allows the system to perform faster similarity comparisons during actual verification, reducing processing time while maintaining high accuracy through the use of pre-computed graph features.
Data Source
AI summary
Embodiments of this application can obtain a request that is initiated at a first resource transfer place and 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; obtain a graph feature associated with at least a second resource transfer place, wherein the graph feature comprises at least one of a resource transfer device graph feature and a target object graph feature; 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 level of the resource transfer object.