Code Authentication System Using Extraction Area Normalization

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In product verification systems, accurately extracting image features from verification targets is challenging due to difficulties in setting consistent shooting directions and sizes, especially when the reference parts like logos and product numbers are not suitable for pre-processing, leading to inaccuracies in determining authenticity.

Innovation Solution

A code authentication system that acquires and processes code images by setting an extraction area within the code image, using pre-processing techniques like normalization, and extracting image features from genuine products for registration and comparison with target images, ensuring accurate authenticity determination by utilizing features like straight sections and right-angle corners.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If image features are extracted from verification target products using reference parts like logos and product numbers, then the system can perform product verification, but the shooting direction and size cannot be set consistently, leading to low measurement precision

Engineering Contradiction:

Improveproduct verification accuracyVSAvoidimage feature extraction accuracy

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The patent extracts and removes the reference part (logo or product number) from the verification process, using it only for positioning and orientation normalization, while the actual authentication features are extracted from the surrounding verification area. This separation allows consistent preprocessing based on reference part position without requiring the reference part itself to be suitable for feature extraction.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The reference part serves as an intermediary element that enables normalization of shooting direction and size, but is not directly used for authentication. It mediates between the variable photographing conditions and the requirement for consistent feature extraction by providing a stable reference for geometric transformation.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Ease of operation

If reference parts like logos and product numbers are used for pre-processing, then shooting direction and size can be normalized, but these reference parts are not suitable for pre-processing, leading to inaccuracies

Engineering Contradiction:

Improvepre-processing capabilityVSAvoidauthenticity determination accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The patent divides the image into distinct functional zones: the reference part (logo or product number) used for positioning and normalization, and the verification area surrounding it used for feature extraction. This segmentation allows each part to serve its optimal function without compromise.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system creates a normalized copy of the verification area based on the position and orientation determined from the reference part, then performs feature extraction on this normalized copy rather than on the original image or the reference part itself, ensuring consistent processing conditions.

Inventive Principle:

Principle #26Copying

3Adaptability or versatility

If image features are extracted without consistent shooting conditions, then the system can process various photographing scenarios, but the determination accuracy decreases

Engineering Contradiction:

Improvephotographing condition flexibilityVSAvoidauthenticity determination accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent implements dynamic normalization that adapts to each photographing scenario by calculating the position and orientation of the reference part in each captured image, then applying geometric transformation specific to that image. This allows the system to handle variable shooting conditions while maintaining consistent feature extraction quality.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS20240386448A1Code authentication system and non-transitory computer readable storage medium

Publication Date: 2024.11.21 DENSO WAVE INC

AI summary

A code image is acquired by photographing a code as an authentication target. An extraction area is set within an area of the code shown in the code image. An image feature extracted from the extraction area is acquired as an extraction image feature. The image feature extracted from the extraction area of the code which is a genuine product is prepared and registered in advance as a registration image feature. An authenticity of the code as the authentication target is determined based on a verification between the extraction image feature and the registration image feature.