Depth-Camera Respiratory Waveforms for Non-Contact Sleep State Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sleep state determination systems are intrusive and inefficient, particularly when monitoring sleep patterns in patients, and lack accuracy in distinguishing between different sleep states.
Innovation Solution
A non-contact system using a depth sensing camera to generate respiratory waveforms from depth images, combined with machine learning models trained on labeled respiratory waveforms to accurately predict sleep states and apnea-hypopnea index without physical contact.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional contact-based sleep monitoring systems are used, then sleep state determination can be achieved, but patient comfort deteriorates and intrusion increases
Solution Approach 1:
The patent replaces mechanical/contact-based sensing systems with optical sensing using depth cameras. The system uses optical depth imaging to capture respiratory movements and derives respiratory waveforms without physical contact, thereby maintaining measurement precision while eliminating the intrusion and discomfort associated with traditional contact-based methods
Solution Approach 2:
The patent introduces an intermediary processing layer that transforms depth camera images into respiratory waveforms through image processing and signal extraction algorithms. This intermediary approach allows indirect measurement of respiratory patterns, achieving accurate sleep state determination without direct physical contact with the patient
2Object-affected harmful factors
If simple monitoring methods are used, then patient comfort is maintained, but sleep state distinction accuracy deteriorates
Solution Approach 1:
The patent changes the measurement parameters from simple presence detection to detailed respiratory waveform analysis. By extracting and analyzing multiple parameters from depth images (respiratory rate, amplitude, patterns), the system achieves high sleep state distinction accuracy while maintaining patient comfort through non-contact monitoring
Solution Approach 2:
The patent transitions from two-dimensional image data to waveform analysis by extracting temporal respiratory patterns. This dimensional transformation from spatial image information to temporal signal analysis enables sophisticated sleep state classification while maintaining the non-intrusive nature of optical sensing
3Productivity
If existing sleep monitoring systems are used, then basic sleep detection is achieved, but apnea-hypopnea index determination accuracy deteriorates
Solution Approach 1:
The patent performs preliminary extraction and labeling of respiratory waveforms during the data collection phase, creating a labeled dataset that captures subtle respiratory patterns. This preliminary action enables the machine learning model to learn and detect apnea-hypopnea events with high accuracy, going beyond basic sleep detection to provide precise clinical metrics
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Implementations described herein disclose a method including determining, based on an image signal received from a camera focused on at least a portion of a patient, a respiratory waveform, receiving an observed sleep signal of the patient temporally corresponding to the respiratory waveform, the observed sleep signal including sleep state labels, labeling various segments of the respiratory waveform using sleep-wake state labels to generate a labeled respiratory waveform, generating an input feature matrix by processing the labeled respiratory waveform, and training a machine learning (ML) model using the input feature matrix.