OSA Diagnosis via Audio Deviation Scores
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for diagnosing respiratory maladies like Obstructive Sleep Apnea (OSA) are labor-intensive, inconvenient, and often inaccurate, relying on identifying characteristic snore segments in patient sounds, which can be computationally expensive and subjective.
Innovation Solution
A method using electronic processors to analyze digital audio signals by identifying epochs, sub-segments, and mel-frequency cepstral coefficients (MFCCs), determining deviation scores from a probability distribution, and applying these to a pre-trained decision machine to generate a malady signal for OSA diagnosis without relying on snore segment identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Polysomnography (PSG) is used for OSA diagnosis, then diagnostic accuracy is improved, but labor intensity and time consumption increase significantly
Solution Approach 1:
The patent replaces the manual mechanical review process of PSG data with an automated acoustic analysis system. The system uses digital signal processing to automatically identify and analyze snore segments, extracting features such as spectral centroid, spectral rolloff, and zero-crossing rate to classify apnea events, thereby eliminating the need for labor-intensive manual review while maintaining diagnostic accuracy
Solution Approach 2:
The patent introduces an automated acoustic analysis algorithm as an intermediary between raw PSG data and diagnostic conclusions. This intermediary system processes the audio signals automatically, identifying characteristic snore patterns and generating diagnostic reports without requiring expert technician intervention, thus resolving the contradiction between accuracy and productivity
2Reliability
If snore segment identification algorithms are used to diagnose OSA, then diagnostic capability is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent divides the lengthy patient sound recording into smaller epochs and further segments each epoch into potential snore segments. By processing the audio signal in manageable segments rather than as a continuous lengthy recording, the system reduces computational complexity while maintaining reliable detection of OSA events throughout the entire recording period
Solution Approach 2:
The patent applies different analysis strategies to different segments of the audio signal. Each segment is analyzed for specific acoustic features characteristic of snoring, and only segments containing snore events are subjected to full OSA analysis. This localized approach reduces overall computational complexity while preserving diagnostic reliability
3Loss of information
If traditional PSG monitoring is used, then comprehensive physiological data is obtained, but patient convenience and accessibility deteriorate
Solution Approach 1:
The patent extracts only the essential acoustic information needed for OSA diagnosis from the full PSG dataset. By focusing specifically on audio signal analysis rather than requiring comprehensive multi-parameter physiological monitoring, the system maintains sufficient diagnostic capability while significantly improving patient convenience and accessibility for population screening
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
A method for diagnosing a malady of a patient from sounds of the patient including the steps of: making a digital recording of the sounds of the patient; processing the digital recording to extract a multiplicity of features for sub-segments of each of a number epochs of the digital recording; determining deviation scores from a probability distribution for each epoch based on extracted multiplicity of features; applying a test vector derived from the deviation scores to a pre-trained decision machine; and presenting a diagnosis of the malady on the basis of an output from said decision machine.