Image Identity Recognition Using Multi-Pattern Look-Alike Verification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image data processing systems struggle to accurately identify the real identity of users within highly look-alike groups, leading to identification failures and reduced accuracy in executing services like payments.

Innovation Solution

Implement a method and apparatus that perform first-type identification on a target object using a look-alike object database, followed by K pattern identification services to confirm the identity, ensuring accurate recognition by outputting a look-alike ID when multiple authentications match.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If single pattern identification is used for look-alike objects, then identification speed is maintained, but identification accuracy deteriorates

Engineering Contradiction:
Improveidentification accuracyVSAvoididentification system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The identification system is segmented into multiple independent pattern identification services (first pattern, second pattern, third pattern, etc.). Each service performs identification using a different pattern or algorithm, and their results are aggregated to make the final determination. This segmentation allows the system to maintain high accuracy through multiple verification points while keeping each individual service relatively simple.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system changes the identification parameters by using multiple different patterns (e.g., different feature extraction methods, different comparison algorithms, or different weightings) to identify the same look-alike object. By varying the identification parameters across multiple services, the system can compensate for the weaknesses of any single pattern and achieve more accurate results.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If multiple pattern identification services are deployed, then identification reliability is improved, but processing time increases

Engineering Contradiction:
Improveidentification reliabilityVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs preliminary action by pre-configuring multiple pattern identification services and their corresponding patterns before actual identification is needed. The services are prepared in advance with their algorithms and parameters, so when identification is required, they can execute quickly without needing to set up or configure anything during the actual processing. This reduces the time penalty of having multiple services.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The multiple pattern identification services operate independently and autonomously, each performing its identification task without requiring manual intervention or coordination. The system automatically collects and aggregates their results, allowing parallel execution that minimizes total processing time while maintaining high reliability through multiple independent verification paths.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4113371B1Image data processing method and apparatus, device, storage medium, and product
Publication Date: 2026.04.01 TENCENT TECHNOLOGY (SHENZHEN) CO LTD
  • EP4113371B1 patent drawingFigure 1
  • EP4113371B1 patent drawingFigure 2
  • EP4113371B1 patent drawingFigure 3

AI summary

Disclosed in embodiments of the present application are an image data processing method and apparatus, a computer device, a computer-readable storage medium, and a computer program product. The method comprises: when performing first-type identity recognition on a target object in an image data stream, if a first recognition result indicates that the target object is a similar object in a similar object database, obtaining similar identity information associated with the similar object, and obtaining, from a similar service configuration database associated with the similar object database, K pattern recognition services configured for the similar identity information; performing second-type identity recognition on the target object in the image data stream by means of the K pattern recognition services to obtain K second recognition results; and if the K second recognition results all indicate that the target object is a similar object, outputting the similar identity information to an application client, such that the application client executes application services on the basis of the similar identity information.