Voice Authentication Obfuscation Against Deepfake Impersonation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice recognition authentication methods are vulnerable to impersonation attacks using deep fake voice models, compromising security in transactions and data access.
Innovation Solution
An authentication system that integrates voice identification with an obfuscation function to alter voice signals before transmission, using scrambling, noise addition, or audio watermarking, and a challenge-response algorithm to generate a unique obfuscation parameter, ensuring secure de-obfuscation at the receiving end.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If voice recognition authentication is used, then authentication convenience is improved, but security against impersonation attacks deteriorates
Solution Approach 1:
The system applies obfuscation to the voice signal before transmission. This preliminary action modifies the voice characteristics in advance so that even if intercepted, the recorded voice cannot be used for impersonation. The obfuscation is applied during the authentication process itself, maintaining convenience while preventing future misuse.
Solution Approach 2:
The system changes the parameters of the voice signal by applying obfuscation functions that modify acoustic characteristics. This transforms the voice signal into a form that is unintelligible for reproduction purposes, thereby preventing deepfake attacks while still allowing legitimate authentication through proper de-obfuscation.
2Measurement precision
If voice samples are collected for authentication, then authentication accuracy is improved, but vulnerability to deepfake attacks increases
Solution Approach 1:
Instead of collecting and storing raw voice samples that could be replicated, the system applies obfuscation during the authentication process itself. This preliminary transformation ensures that the voice characteristics are modified in real-time, making them unusable for creating deepfakes while still enabling accurate authentication through the de-obfuscation process.
Solution Approach 2:
The system takes preliminary anti-action by obfuscating the voice signal before it can be recorded or analyzed for deepfake creation. This preemptive measure counteracts the potential harmful effect of voice sample collection by ensuring that any recorded signal is already distorted and unusable for impersonation purposes.
3Reliability
If obfuscation is applied to voice signals, then security is improved, but system complexity increases
Solution Approach 1:
The system introduces an obfuscation function as an intermediary between the voice signal and the authentication process. This intermediary component adds security by transforming the voice signal, while the overall complexity remains manageable because the obfuscation and de-obfuscation operations are computationally efficient and can be integrated into existing authentication systems.
Data Source
Figure 1~2
Figure 3~5
Figure 6~8
AI summary
A voice authentication method (100), for authenticating a user by voice at an identification terminal (10), comprising the steps of: - for an authentication instance, acquiring (102) a voice signal representative of the user's voice by means of a user terminal (1); - generating (103), by means of the user terminal (1), at least one obfuscation parameter (Kcr) of the voice signal, wherein the obfuscation parameter (Kcr) is a parameter specifically generated for said authentication instance; - processing (104), with the user terminal (1), the voice signal by applying an obfuscation function to the voice signal to generate an obfuscated voice signal, wherein said obfuscation parameter (Kcr) controls the obfuscation function; - transmitting (105), with the user terminal (1), the obfuscated voice signal to the identification terminal (10); - generating (106), at the identification terminal, said obfuscation parameter (Kcr); - processing (107), at the identification terminal (10), the obfuscated signal to apply (107) a de-obfuscation function to the obfuscated voice signal and generate a de-obfuscated voice signal, wherein the obfuscation parameter (Kcr) controls the de-obfuscation function; - analyzing (108), at the identification terminal (10), the de-obfuscated voice signal; - determining (109), based on the outcome of said step of analyzing (108), whether the de-obfuscated voice signal is authentic or not.