Voice Biometric Audio Segmentation and Noise Removal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice biometrics face challenges in obtaining a clear user voice sample due to background noise and inconvenience in finding a quiet environment, especially during phone calls where multiple voices and environmental sounds interfere with audio recordings.
Innovation Solution
A biometric system uses machine learning to differentiate between user and agent voices in audio recordings from phone calls by generating signature vectors and comparing them using a distance metric to remove non-matching segments, thereby isolating the user's voice for authentication purposes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If voice biometrics are used to authenticate users during phone calls, then user authentication can be performed conveniently without requiring a quiet environment, but background noise and multiple voices interfere with the audio recording making it difficult to obtain a clear user voice sample
Solution Approach 1:
The audio recording is divided into multiple segments, and each segment is analyzed individually to identify portions containing only the user's voice. This segmentation allows the system to separate the user's voice from background noise and other speakers by processing discrete time windows of the audio signal.
Solution Approach 2:
The system extracts the user's voice signal from the mixed audio recording by comparing audio segments against a reference voice print. Segments that match the user's voice characteristics are extracted and isolated, while portions containing background noise or other speakers are removed.
2Quantity of substance
If the entire audio recording is used for biometric authentication, then more voice data is available for analysis, but non-matching portions including background noise and other voices reduce the accuracy of the biometric sample
Solution Approach 1:
The system uses a feedback mechanism where each audio segment is compared against the user's voice print, and the comparison result determines whether that segment is included in the final biometric sample. This feedback loop ensures that only segments matching the user's voice characteristics are retained, automatically filtering out noise and other speakers.
Solution Approach 2:
The system changes the parameter of audio segment inclusion based on similarity comparison results. Segments are dynamically selected or excluded from the biometric sample based on whether they meet a similarity threshold with the reference voice print, optimizing the quality of the final biometric data.
3Measurement precision
If machine learning models are used to generate voice signatures and compare audio segments, then the system can automatically differentiate user voice from noise and other voices, but the processing complexity and computational requirements increase
Solution Approach 1:
The system performs preliminary actions by pre-processing the audio recording into segments before comparison. This segmentation and preliminary analysis simplifies the subsequent machine learning comparison task, as the model only needs to evaluate discrete segments against the voice print rather than processing the entire continuous audio stream at once.
Data Source
AI summary
Methods and systems are disclosed herein for improving the quality of audio for use in a biometric. A biometric system may use machine learning to determine whether audio or a portion of the audio should be used as a biometric for a user. A sample of the user's voice may be used to generate a voice signature of the user. Portions of the audio that do not meet a similarity threshold when compared with the voice signature may be removed from the audio. Additionally or alternatively, interfering noises may be detected and removed from the audio to improve the quality of a voice biometric generated from the audio.