Speaker Authentication Using Delimiter-Separated Verification Strings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speaker recognition systems for user authentication face challenges such as coarticulation effects, spoken ambiguities, high costs, language dependency, and reduced usability due to the need for long digit strings, which degrade accuracy and user verification.
Innovation Solution
A method and system that uses a verification string with delimiter characters to minimize coarticulation impact, employing a lightweight alphanumeric character recognizer based on hidden Markov models or dynamic time warping, allowing for improved accuracy and easier language portability by separating alphanumeric characters with delimiters like dashes, periods, or commas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a long digit string is used for user authentication, then the system can achieve reasonable accuracy in speaker recognition, but the usability is reduced
Solution Approach 1:
The verification string is segmented into individual character elements separated by delimiters (e.g., '1-2-3-4-5' instead of '12345'). This segmentation allows the system to process and verify each character independently, maintaining recognition accuracy while reducing the perceived length and complexity for the user, thereby improving usability.
2Reliability
If random text is used for verification, then playback attacks can be prevented, but coarticulation effects and spoken ambiguities degrade authentication accuracy
Solution Approach 1:
The system applies different processing qualities to different parts of the verification string. Delimiter characters (e.g., dashes, periods) are processed with specialized recognition models that account for their unique acoustic properties, while alphanumeric characters use separate models. This local quality approach reduces coarticulation effects by treating each character type according to its specific characteristics, thereby improving authentication accuracy while maintaining security.
3Extent of automation
If traditional speaker recognition systems are implemented, then user authentication can be performed, but the systems are expensive and language-dependent
Solution Approach 1:
The system uses a universal set of delimiter characters (dashes, periods, commas) that can be recognized across multiple languages and regions. The lightweight alphanumeric character recognizer is designed to be language-agnostic, allowing the same system to function in different linguistic contexts without requiring language-specific models, thereby reducing complexity and cost while maintaining authentication capability.
4Length of stationary object
If alphanumeric characters are presented without delimiters, then the verification string can be shorter, but coarticulation effects increase and reduce recognition accuracy
Solution Approach 1:
Delimiter characters serve as intermediaries between alphanumeric characters in the verification string. These delimiters act as acoustic boundaries that prevent coarticulation effects from propagating between adjacent alphanumeric characters. The system recognizes these intermediary delimiters using specialized models, allowing the overall verification string to remain relatively short while maintaining high recognition accuracy through the mediating role of the delimiters.
Data Source
AI summary
Methods, computing systems and computer program products implement embodiments of the present invention that include defining a verification string including a sequence of verification characters and a delimiter character between each sequential pair of the verification characters, the delimiter character being different from the verification characters. The verification string to a user, and upon receiving, from the user, a series of verification vocal inputs in response to presenting the verification string, a set of verification features from each of the verification vocal inputs are computed so as to generate sets of verification features. A one-to-one correspondence is established between each of the verification vocal inputs and each of the verification characters, and the user is authenticated based on the verification vocal inputs and their corresponding sets of verification features.