Biometric Authentication via Acoustic Head Response Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Acoustic characteristics acquired by wearable devices, such as earphones, can change due to variations in wearing state, affecting authentication accuracy, necessitating a robust feature extraction method.
Innovation Solution
An information processing device and method that acquires acoustic characteristics in the frequency domain and generates a first frequency response function using a rational polynomial equation, incorporating terms for peak and notch features, to extract robust biometric authentication features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If acoustic characteristics are acquired based on sound wave propagation in user's head for biometric authentication, then authentication function is enabled, but authentication accuracy deteriorates due to changes in acoustic characteristics from wearing state variations
Solution Approach 1:
The patent transforms the acoustic characteristics from time-domain signal to frequency-domain representation using Fourier transform. This parameter transformation allows the system to identify and extract stable frequency components (formants) that are less sensitive to wearing state variations, thereby improving measurement precision while maintaining authentication functionality
Solution Approach 2:
The patent extracts specific feature quantities (formant frequencies F1, F2, F3 and their bandwidths) from the complete acoustic characteristic spectrum. By isolating these particular frequency components that represent vocal tract resonances, the system obtains authentication features that are robust against wearing state changes, resolving the contradiction between enabling authentication and maintaining accuracy
Data Source
AI summary
Provided is an information processing device comprising an acquisition unit configured to acquire acoustic characteristics in a frequency domain based on a sound wave propagating in a head of a user and an extracting unit configured to generate a first frequency response function having a rational polynomial equation to extract feature quantity used for biometric authentication of the user based on the first frequency response function, the rational polynomial equation including terms indicating feature of a peak of the acoustic characteristics in a denominator and terms indicating feature of a notch of the acoustic characteristics in a numerator.