Distributed Face Library for Multi-Document Risk Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In risk control systems, especially in credit businesses, accurate risk assessment is hindered by the inability to uniquely identify users in multi-document scenarios, leading to potential losses due to malicious activities, as existing systems struggle to differentiate between real and fake identities and handle diverse documents effectively.
Innovation Solution
A method and system utilizing a distributed face library that retrieves and matches living face images based on a preset retrieval strategy, dividing the library into sub-libraries by common attributes or data features to enhance search efficiency and accuracy, thereby improving risk control capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If document recognition is used to identify users, then the system can process various document types, but it cannot uniquely identify users in multi-document scenarios
Solution Approach 1:
The patent divides the face library into multiple sub-libraries based on different document types (e.g., driver's license, passport, social security card). Each sub-library stores face images associated with a specific document type, allowing the system to segment the identification process by document category while maintaining unique user identification across all document types
Solution Approach 2:
The patent introduces a face image as an intermediary element that links different document types to a unique user identity. Instead of relying solely on document numbers which vary across document types, the system uses face recognition as a mediator to establish consistent user identification across diverse documents
2Quantity of substance
If a centralized face library is used for face matching, then all face images can be stored in one place, but retrieval and matching speed decreases
Solution Approach 1:
The patent segments the centralized face library into multiple distributed sub-libraries, each storing a portion of the total face images. This segmentation allows parallel processing during face matching operations, where the system can simultaneously search multiple sub-libraries, thereby maintaining comprehensive storage capacity while significantly improving retrieval and matching speed
3Reliability
If face matching is performed across all documents, then accurate risk identification can be achieved, but system complexity increases
Solution Approach 1:
The patent organizes the face library into structured sub-libraries based on document types, creating a hierarchical system that simplifies the matching process. Instead of performing brute-force comparisons across all documents, the system first identifies the document type and queries the corresponding sub-library, thereby maintaining high risk identification accuracy while reducing system complexity through organized data structure
Solution Approach 2:
The patent applies different matching strategies and criteria to different document types by organizing face images into type-specific sub-libraries. Each sub-library can be optimized with local quality parameters appropriate for its document category, allowing the system to maintain high reliability across diverse documents without requiring a single complex unified approach
Data Source
AI summary
Methods and systems for controlling a risk based on a distributed face library are provided. The method may be implemented by a processor and may include obtaining a living face image and generating a face retrieval demand signal; calling the distributed face library based on the face retrieval demand signal, the distributed face library including a plurality of sub-libraries; matching the living face image through the distributed face library based on a preset retrieval strategy; in response to a determination that the living face image is not matched, returning a first risk result; or in response to a determination that the living face image is matched, returning a second risk result. The living face image may be retrieved and matched quickly through the method for controlling a risk based on a distributed face library, which may quickly recognize a user identity of a unique identifier in a multi-document scenario, thereby improving a risk control capability of the system.