Voice-Based Oral Function Evaluation with S/N-Adaptive Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for evaluating oral function, such as those requiring appliances or visual inspection, cause discomfort and may overlook deterioration, leading to undernutrition and aspiration risks, while voice-based evaluations lack accuracy due to noise interference.
Innovation Solution
An oral function evaluation device and method that analyzes voice data to determine a signal-to-noise ratio, selecting appropriate estimating equations based on this ratio to accurately assess oral function deterioration, using features like sound pressure and formant frequencies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If voice data is collected for oral function evaluation, then evaluation convenience is improved, but measurement precision deteriorates due to noise interference
Solution Approach 1:
The patent segments the evaluation process into two distinct paths based on S/N ratio thresholds. When noise is low (S/N ratio above threshold), the system uses a comprehensive evaluation model including sound pressure features. When noise is high (S/N ratio at or below threshold), it switches to an alternative evaluation model that excludes sound pressure features, thereby adapting to different environmental conditions and maintaining evaluation accuracy.
Solution Approach 2:
The patent dynamically adjusts the evaluation methodology based on real-time noise conditions. The system calculates the S/N ratio and dynamically selects which estimating equation to use, transitioning between different feature sets (including or excluding sound pressure) depending on the acoustic environment, thus optimizing measurement precision under varying operational conditions.
2Measurement precision
If sound pressure features are used in estimation, then measurement precision is improved, but reliability deteriorates in noisy environments
Solution Approach 1:
The patent changes the parameters used in estimation based on noise conditions. When the S/N ratio indicates a noisy environment, the system excludes sound pressure features from the estimation parameters, using only formant frequency features instead. This parameter adaptation ensures reliable evaluation even when acoustic conditions are unfavorable.
Solution Approach 2:
The S/N ratio calculation acts as an intermediary that mediates between the voice data and the estimation process. It evaluates the quality of the acoustic environment and determines which features should be trusted, thereby protecting the reliability of the evaluation by filtering out unreliable sound pressure measurements in noisy conditions.
3Measurement precision
If appliances are used for evaluation, then measurement precision is improved, but ease of operation deteriorates due to discomfort
Solution Approach 1:
The patent replaces mechanical appliances (such as pharynx sensors that require physical attachment) with an acoustic field-based evaluation system using microphones and voice analysis. This substitution eliminates the need for physical contact with the patient's body, improving comfort while maintaining evaluation capability through voice-based feature extraction.
Data Source
AI summary
An oral function evaluation device includes: an obtainer that obtains voice data; an extractor that extracts a feature from the voice data; an S/N ratio calculator that calculates a signal-to-noise (S/N) ratio of a second average intensity of a sound collected in a period when a voice is uttered to a first average intensity of a sound collected in a period when no voice is uttered; a determiner that determines an estimating equation; and an evaluator that evaluates the deterioration state of oral function through assessment using an oral function evaluation indicator, based on the estimating equation determined and the feature extracted. As the estimating equation, the determiner determines: a first estimating equation that includes a sound-pressure-related feature when the S/N ratio is greater than a first threshold; and a second estimating equation that does not include the sound-pressure-related feature when the S/N ratio is the first threshold or less.