Sleep Quality Estimation via Audio Signal Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for sleep diagnosis, such as Polysomnography, are invasive, costly, and not suitable for mass population due to their requirement for specialized equipment and trained technicians, necessitating the development of non-invasive and cost-effective methods for estimating sleep stages and breathing patterns.
Innovation Solution
A system and method using audio signal processing to determine sleep quality parameters by segmenting audio signals into epochs, extracting feature parameters, and applying machine learning models to classify sleep stages and breathing patterns, including the use of breathing detection and noise reduction algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Polysomnography is used for sleep diagnosis, then measurement precision is improved, but device complexity and cost increase
Solution Approach 1:
The patent extracts only the essential acoustic information needed for sleep stage detection from the complex polysomnography system. By using a single non-contact microphone to capture breathing sounds and eliminating the need for multiple contact-based sensors (EEG, EOG, EMG electrodes), the system achieves acceptable sleep stage detection accuracy with dramatically reduced device complexity
Solution Approach 2:
The patent replaces the mechanical contact-based sensor system with an acoustic field-based detection system. Instead of using physical electrodes that require skin contact and complex wiring, the system uses acoustic waves captured by a non-contact microphone to detect breathing patterns and infer sleep stages, thereby reducing device complexity while maintaining measurement capability
2Measurement precision
If Polysomnography is used for sleep diagnosis, then measurement precision is improved, but ease of operation deteriorates
Solution Approach 1:
The patent implements automated sleep stage detection using machine learning algorithms that process acoustic signals independently without requiring human intervention. The system automatically segments sleep periods, extracts breathing features, classifies sleep stages, and generates reports, eliminating the need for trained technicians to manually operate complex equipment and interpret results
Solution Approach 2:
The patent introduces an automated computational intermediary (machine learning model) that bridges the gap between raw acoustic signals and sleep stage classification. This intermediary automatically performs the complex analysis that previously required trained technicians, thereby improving ease of operation while maintaining measurement precision
3Measurement precision
If contact-based sensors are used in Polysomnography, then measurement precision is improved, but object-affected harmful factors increase
Solution Approach 1:
The patent replaces contact-based mechanical sensors with a non-contact acoustic detection system. By using sound waves to capture breathing information without physical contact, the system eliminates the harmful effects of sensors on sleep quality (discomfort, movement restriction, skin irritation) while still obtaining accurate physiological data for sleep stage detection
4Ease of operation
If non-contact audio methods are used for sleep estimation, then ease of operation is improved, but measurement precision deteriorates
Solution Approach 1:
The patent changes the analytical parameters and processing methods for acoustic signals to extract more informative features. By using advanced signal processing techniques (spectral analysis, breathing pattern recognition, feature extraction from acoustic signals) and machine learning classification, the system achieves measurement precision comparable to contact-based methods while maintaining the ease of non-contact operation
Data Source
AI summary
The present invention relates to a system and method for determining sleep quality parameters according to audio analyses, comprising: obtaining an audio recorded signal comprising sleep sounds of a subject; segmenting the signal into epochs; generating a feature vector for each epoch, wherein each of said feature vectors comprises one or more feature parameters that are associated with a particular characteristic of the signal and that are calculated according to the epoch signal or according to a signal generated from the epoch signal; inputting the generated feature vectors into a machine learning classifier and applying a preformed classifying model on the feature vectors that outputs a probabilities vector for each epoch, wherein each of the probabilities vectors comprises the probabilities of the epoch being each of the sleep quality parameters; inputting the probabilities vectors for each epoch into a machine learning time series model and applying a preformed sleep quality time series pattern function on said probabilities vectors that outputs an enhanced probabilities vector for each epoch; determining a final sleep quality parameter for each epoch by calculating the most probable sleep quality parameters sequence.