Handwritten Signature Authentication Using Behavioral and Reconstructed Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing handwritten signature authentication technologies suffer from low recognition and accuracy rates, and there is a need for improved methods to securely verify user identity in online environments where face-to-face verification is not possible, particularly when mobile communication terminals are lost or information is leaked.
Innovation Solution
A system and method for handwritten signature authentication that compares pre-enrolled handwritten signature images with tracked and reconstructed images, utilizing behavioral characteristics data to enhance authentication by normalizing successfully authenticated images and behavioral data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional handwritten signature authentication is used, then the authentication process is simple, but the recognition rate and accuracy rate are low
Solution Approach 1:
The patent segments the signature verification process into multiple independent comparison components: (1) pre-enrolled signature image vs. tracked signature image comparison, (2) behavioral characteristics comparison, and (3) reconstructed image vs. pre-enrolled image comparison. Each component operates independently and contributes to the overall authentication decision, thereby improving accuracy without creating a monolithic complex system.
Solution Approach 2:
The patent adds a new dimension to traditional signature authentication by incorporating behavioral characteristics (pressure, speed, acceleration, direction) alongside the visual image data. This multi-dimensional approach transforms the authentication from purely 2D image matching to a multi-parameter verification system, significantly improving recognition accuracy.
2Reliability
If multiple authentication technologies are combined, then user security is improved, but the authentication process becomes more complex
Solution Approach 1:
The patent merges multiple authentication elements (image data, pressure data, speed data, acceleration data, direction data) into a unified handwritten signature authentication system. These elements are collected simultaneously during a single signature input action and processed together through integrated comparison units, achieving multi-factor authentication without requiring separate authentication steps or systems.
Solution Approach 2:
The authentication system is designed to handle multiple types of data (image, pressure, motion characteristics) through a universal processing framework. The same enrollment and verification流程apply to all data types, and the system can adapt to different input devices (touchscreen, stylus, scanner) while maintaining consistent authentication logic.
3Measurement precision
If behavioral characteristics data is collected and analyzed, then authentication accuracy is improved, but data processing requirements increase
Solution Approach 1:
The patent extracts only the essential behavioral characteristics (pressure, speed, acceleration, direction) from the raw touch input data, rather than processing all available data. These key parameters are extracted and stored as behavioral characteristics data, which are then used for comparison during verification, reducing the data volume while maintaining authentication accuracy.
Solution Approach 2:
The system performs preliminary processing of behavioral characteristics data during the enrollment phase, where typical signature patterns and characteristics are established and stored as reference data. During verification, only the extracted behavioral parameters need to be compared against these pre-established patterns, reducing real-time processing requirements.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to a manual signature authentication system and method and, more specifically, to a manual signature authentication system for performing manual signature authentication by comparing at least one among a pre-registered manual signature image for a user's manual signature, a tracking manual signature image tracked during a manual signature, a manual signature action characteristic collected during the manual signature, and a manual signature image reconstructed by the manual signature action characteristic, and a method therefor.