Respiratory Cycle Extraction from Auditory Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for monitoring respiratory illnesses, such as cystic fibrosis, rely on frequent CT scans that expose patients to harmful radiation, and digital auscultation is subjective and prone to noise interference, making it difficult to accurately detect lung abnormalities like crackles.
Innovation Solution
A computer-implemented system that de-noises lung sounds using a neural network and wavelet transform to identify and count respiratory abnormalities, reducing radiation exposure and improving detection precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frequent CT scans are used to monitor respiratory illnesses, then monitoring accuracy is improved, but patient radiation exposure increases
Solution Approach 1:
The patent replaces the mechanical/radiological imaging system (CT scans) with an acoustic sensing system (digital auscultation). Instead of using X-rays to image lung structure, the system uses microphones to capture lung sounds, substituting a harmful physical method with a non-invasive acoustic measurement approach that provides functional respiratory monitoring.
Solution Approach 2:
The patent creates an acoustic copy or representation of respiratory function through recorded lung sounds. Rather than repeatedly exposing patients to radiation for direct imaging, the system captures acoustic signatures of breathing that can be analyzed multiple times without additional radiation exposure, effectively copying the respiratory information in a safe manner.
2Object-affected harmful factors
If digital auscultation is used to monitor lung sounds, then radiation exposure is reduced, but detection precision deteriorates due to noise interference
Solution Approach 1:
The patent extracts the relevant respiratory signal from the noisy auditory signal by identifying and isolating specific acoustic features characteristic of lung sounds. The system separates the useful respiratory information from background noise and motion artifacts through signal processing techniques that extract only the medically relevant components.
Solution Approach 2:
The patent transforms the auditory signal from the time domain to the frequency domain using Fourier transform, changing the representation parameters of the signal. This parameter transformation allows noise filtering by operating in the frequency domain where respiratory signals and noise can be distinguished and separated through spectral analysis.
3Ease of operation
If human doctors manually analyze lung sounds, then clinical judgment is applied, but measurement precision deteriorates due to inability to count multiple crackles
Solution Approach 1:
The patent replaces the human auditory system and manual counting process with an automated computational analysis system. The computer-based system can process and count multiple crackles within a respiratory cycle that would be imperceptible to human ears, substituting biological limitations with computational capabilities that provide precise quantitative measurements.
Solution Approach 2:
The patent creates a digital representation or copy of the lung sounds that can be analyzed frame-by-frame to count individual crackles. By capturing the acoustic signal digitally and analyzing it through computational methods, the system creates a measurable copy of the respiratory sounds that enables precise counting beyond human perceptual capabilities.
4Measurement precision
If noise filtering is applied to lung sounds, then detection precision is improved, but loss of information may occur about relevant signals
Solution Approach 1:
The patent transforms the signal to the frequency domain using Fourier transform, changing from time-domain analysis to frequency-domain analysis. This parameter change allows selective filtering of noise frequencies while preserving the frequency bands containing respiratory signals, enabling noise reduction without losing medically relevant information through intelligent frequency-based discrimination.
Data Source
AI summary
Aspects disclosed herein disclose a system and method for improving digital stethoscopes and their application and operation. A first method de-noises an auditory signal. A second decomposes an auditory signal into sub-components. A third method extracts a respiratory cycle from the auditory signal. A fourth method counts respiratory abnormalities based on the respiratory cycle.