Ensemble Respiratory Screening From Voice and Cough Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Clinical diagnosis of COVID-19 is time-consuming and expensive, especially in areas with limited resources, and existing computer-implemented methods often rely on a single channel of information, leading to lower accuracy and specificity in detecting respiratory infections.
Innovation Solution
Implementing ensemble machine-learning models that combine multiple channels of input data, including audio and image analysis, using smartphones or other devices to gather data from users, and employing a variety of preprocessing techniques to enhance accuracy and specificity in detecting COVID-19 and other respiratory syndromes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If clinical diagnosis methods are used for COVID-19 detection, then diagnostic accuracy is improved, but time consumption and cost increase significantly
Solution Approach 1:
The patent replaces complex mechanical clinical diagnostic procedures with an acoustic analysis system that processes voice and cough signals through machine learning models. The system substitutes physical examination and laboratory testing with computational acoustic feature extraction and classification, achieving rapid automated diagnosis without requiring clinical infrastructure.
Solution Approach 2:
The patent introduces acoustic signals (voice and cough recordings) as an intermediary medium between the patient and the diagnostic system. These acoustic recordings serve as proxies that capture respiratory system information, enabling indirect but accurate diagnosis through machine learning analysis of sound patterns rather than direct clinical examination.
2Device complexity
If single-channel information methods are used for respiratory infection detection, then system complexity is reduced, but detection accuracy and specificity decrease
Solution Approach 1:
The patent merges multiple independent acoustic analysis channels (voice analysis, cough analysis, and their combination) into a unified diagnostic system. Each channel processes different aspects of respiratory sounds independently, then their results are combined through ensemble machine learning methods, achieving superior accuracy that exceeds any single channel alone while maintaining modular system architecture.
Solution Approach 2:
The patent segments the diagnostic system into distinct functional modules: voice recording module, cough recording module, acoustic feature extraction module, and classification module. Each segment handles specific processing tasks independently, allowing the complex multi-channel system to be built from manageable components that can be developed and validated separately before integration.
Data Source
AI summary
Provided is a process including: obtaining, with one or more processors, a set of data comprising a plurality of patient records, selecting a subset of the plurality of parameters for inputs into a machine learning system, generating a classifier using the machine learning system based on the training data and the subset of the plurality of parameters for inputs; receiving, with one or more processors, patient record of a first user; performing an analysis, with one or more processors, to identify acoustic measures from a voice sample of the first user.