Face Authentication Apparatus Dynamic Registration Threshold

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing personal authentication methods using face images are prone to erroneous recognition due to changes in face orientation, lighting conditions, and environmental variations, leading to increased storage burdens and processing times, as well as erroneous authentication results from displacement of small areas in face images.

Innovation Solution

An authentication apparatus and method that detects specific areas in images, extracts feature information, and compares it with stored data, determining whether to register new feature information based on a dynamic registration threshold value that adjusts according to the number of registered data items, thereby reducing erroneous authentication.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If auxiliary feature value data items are stored for each different shooting condition including different environments, then personal authentication accuracy is improved, but storage burden and processing time increase significantly

Engineering Contradiction:

Improvepersonal authentication accuracyVSAvoidamount of dictionary data

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments the shooting conditions into two categories: capture environment (lighting, background) and camera environment (focal length, aperture). By dividing and storing feature values according to these segmented conditions separately, the system achieves comprehensive coverage without creating an unmanageably large dataset, thus resolving the contradiction between authentication accuracy and data storage burden

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent changes the parameters of feature value storage by organizing data according to specific shooting condition parameters (capture environment and camera environment) rather than storing all possible variations. This parameter-based organization allows efficient storage and retrieval, improving authentication accuracy while controlling data quantity

Inventive Principle:

Principle #35Parameter changes

2Reliability

If combined part features are generated from multiple registered patterns, then authentication robustness is improved, but degree of similarity changes depending on number of registered patterns causing erroneous authentication

Engineering Contradiction:

Improveauthentication robustnessVSAvoiddegree of similarity accuracy

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The patent applies dynamics by making the registration threshold value changeable based on the number of registered pattern groups. As more patterns are registered, the threshold adapts accordingly, preventing erroneous authentication while maintaining robustness. This dynamic adjustment resolves the contradiction between authentication reliability and similarity measurement precision

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system implements feedback by using the registration threshold value to control whether new feature values are registered or rejected. The threshold provides a feedback mechanism that prevents erroneous authentication by comparing new data against established patterns, thus maintaining both reliability and precision in similarity measurement

Inventive Principle:

Principle #23Feedback

3Adaptability or versatility

If face images are stored under various shooting conditions, then coverage of different environments is improved, but user operation complexity increases as users must know shooting conditions

Engineering Contradiction:

Improvecoverage of shooting conditionsVSAvoiduser operation simplicity

Core Design Contradiction:

Adaptability or versatilityVSEase of operation

Solution Approach 1:

The patent enables self-service by allowing the system to automatically capture and store feature values under various shooting conditions without requiring user knowledge or input about the conditions. The system autonomously adapts to different environments and camera settings, achieving comprehensive coverage while maintaining ease of operation

Inventive Principle:

Principle #25Self-service

Data Source

PatentUS9152846B2Authentication apparatus that authenticates object, authentication method, and storage medium

Publication Date: 2015.10.06 CANON KK

AI summary

An authentication apparatus capable of reducing erroneous authentication. A face detection section detects a face area of an object from an image. A feature information extraction processor extracts feature information (image data) indicative of a feature of the object. An authentication determination section performs authentication by comparing registered image data and feature information of a specific object. A registration information processor determines, when one of objects associated with registered image data items is selected as an object to which the feature information of the specific object is to be added, whether or not to additionally register the feature information of the specific object as image data for the selected object, according to a degree of similarity between image data of the selected object and the feature information of the specific object.