Voice Password Authentication Using Word Segmentation and Intermediary Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current voice recognition-based password authentication systems face issues with password leakage and false recognition, particularly when using voice inputs, and raise privacy concerns due to the storage of voice content, leading to limitations in security and usability.
Innovation Solution
An electronic apparatus that selects and transmits multiple words corresponding to characters, performs password authentication based on user voice inputs, and combines characters based on reception order and time differences, enhancing security and usability by preventing password leakage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a user directly utters a password through voice recognition, then password authentication can be performed, but the password may be leaked and recognition errors may occur
Solution Approach 1:
The password is segmented into multiple character components, each represented by a separate word from a predefined set. Instead of recognizing a continuous password utterance, the system identifies individual words corresponding to individual characters, reducing recognition errors and improving security by preventing direct password exposure.
Solution Approach 2:
A predefined set of words serves as an intermediary layer between the user's voice input and the actual password. The user utters words from this set rather than the password itself, and the system maps these words to corresponding characters through a predetermined relationship, preventing password leakage while maintaining authentication functionality.
2Reliability
If sentence-based passwords are used instead of common passwords, then security is enhanced, but password leakage risk increases when uttered
Solution Approach 1:
The system introduces a word set as an intermediary that masks the actual password. Users utter words from this predefined set rather than the password itself, and the system reconstructs the password by mapping words to characters based on a predetermined relationship. This eliminates password leakage risk during voice transmission while maintaining high security through unique sentence-based passwords.
3Ease of operation
If voice content is stored in a server for recognition, then authentication can be performed, but privacy concerns arise due to storage of password content
Solution Approach 1:
The predefined word set acts as an intermediary that protects user privacy. Instead of storing and processing actual password content, the system stores and transmits only the words from the predefined set that the user utters. The mapping relationship between words and password characters is maintained securely, enabling authentication functionality while preventing exposure of sensitive password information.
4Reliability
If multiple words corresponding to characters are selected and transmitted, then security is improved by preventing password leakage, but device complexity increases
Solution Approach 1:
The password authentication process is segmented into discrete steps: selecting individual words from a predefined set, transmitting these words separately, receiving voice inputs for each word, and mapping words to corresponding characters. This segmentation simplifies the overall system architecture by breaking down the complex task of secure password authentication into manageable, independent operations.
Data Source
Figure 1~2a
Figure 2b~3a
Figure 3b~3c
AI summary
An electronic apparatus is disclosed. The electronic apparatus includes a communicator; and a processor configured to select a plurality of different words respectively corresponding to a plurality of characters based on occurrence of a password input event, and to transmit the plurality of selected words to a terminal device through the communicator, and based on a user voice including at least one of the plurality of selected words being received, perform password authentication based on a character corresponding to the at least one of the plurality of selected words.