Abstract:Continuous cough monitors can greatly aid doctors in home monitoring and treatment of respiratory diseases. Although many algorithms have been proposed, they still face limitations in data privacy and short-term monitoring. Edge-AI offers a promising solution by processing privacy-sensitive data near the source, but challenges arise in deploying resource-intensive algorithms on constrained devices. From a suitable selection of audio and kinematic signals, our methodology aims at the optimal selection of features via Recursive Feature Elimination with Cross-Validation (RFECV), which exploits the explainability of the selected XGB model. Additionally, it analyzes the use of Mel spectrogram features, instead of the more common MFCC. Moreover, a set of hyperparameters for a multimodal implementation of the classifier is explored. Finally, it evaluates the performance based on clinically relevant event-based metrics. We apply our methodology to develop Cough-E, an energy-efficient, multimodal and edge AI cough detection algorithm. It exploits audio and kinematic data in two distinct classifiers, jointly cooperating for a balanced energy and performance trade-off. We demonstrate that our algorithm can be executed in real-time on an ARM Cortex M33 microcontroller. Cough-E achieves a 70.56\% energy saving when compared to the audio-only approach, at the cost of a 1.26\% relative performance drop, resulting in a 0.78 F1-score. Both Cough-E and the edge-aware model optimization methodology are publicly available as open-source code. This approach demonstrates the benefits of the proposed hardware-aware methodology to enable privacy-preserving cough monitors on the edge, paving the way to efficient cough monitoring.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to develop a cough detection algorithm that can achieve real - time, high - efficiency, and privacy - protection, which is used to monitor and record patients' coughs on edge devices (such as microcontrollers). Specifically, this research aims to overcome several key problems existing in the current cough monitoring systems: 1. **Data privacy issues**: Many existing cough monitoring systems need to transmit sensitive data such as audio to the cloud or local computers for processing, which may lead to privacy leakage. The method proposed in this paper protects user privacy by directly processing data on edge devices and avoiding data transmission. 2. **Limitations of short - term monitoring**: Traditional cough monitoring methods are usually only capable of short - term monitoring and cannot capture the long - term changes of cough conditions. The Cough - E algorithm proposed in this paper can continuously monitor patients' coughs for a long time and provide more accurate health information. 3. **Performance optimization on resource - limited devices**: Edge devices (such as ARM Cortex M33 microcontrollers) have limited computing resources and it is difficult to run complex deep - learning models. Through the hardware - aware design method, this paper significantly reduces energy consumption and computing requirements while ensuring detection performance, enabling the algorithm to run in real - time on resource - limited devices. 4. **Multi - modal data fusion**: This paper uses the combination of audio and motion signals (such as accelerometer data) to improve the accuracy of cough detection. Through the joint use of multi - modal data, the Cough - E algorithm not only improves the detection accuracy but also reduces the false alarm rate of single - modal systems. 5. **Clinically relevant performance evaluation**: Traditional cough detection algorithms mainly focus on the sensitivity and specificity at the segment level, but these indicators cannot fully reflect the detailed information required clinically. This paper proposes event - based evaluation indicators, such as the exact number of coughs, the number and duration of cough bursts, etc., to better meet clinical needs. In summary, by designing a hardware - aware multi - modal cough detection algorithm Cough - E, this paper solves the problems of existing systems in data privacy, long - term monitoring, performance optimization on resource - limited devices, and clinically relevant performance evaluation, providing a new solution for achieving efficient edge AI cough monitoring.

Cough-E: A multimodal, privacy-preserving cough detection algorithm for the edge

HearCough: Enabling continuous cough event detection on edge computing hearables

A Machine Hearing System for Robust Cough Detection Based on a High-Level Representation of Band-Specific Audio Features

EarCough: Enabling Continuous Subject Cough Event Detection on Hearables

Design of an Efficient CNN-based Cough Detection System on Lightweight FPGA

Automatic Non-Invasive Cough Detection based on Accelerometer and Audio Signals

Cough Detection Using Acceleration Signals and Deep Learning Techniques

How to Count Coughs: An Event-Based Framework for Evaluating Automatic Cough Detection Algorithm Performance

Deep Neural Network Based Cough Detection Using Bed-Mounted Accelerometer Measurements

Cough Detection Using Selected Informative Features from Audio Signals

An automatic cough counting method and system construction for portable devices

CoughTrigger: Earbuds IMU Based Cough Detection Activator Using An Energy-efficient Sensitivity-prioritized Time Series Classifier

Leveraging cough sounds to optimize chest x-ray usage in low-resource settings

Cough Detection from Acoustic signals for patient monitoring system

Audio and Contact Microphones for Cough Detection

EDAC: Efficient Deployment of Audio Classification Models For COVID-19 Detection

Wearable RF Near-Field Cough Monitoring by Frequency-Time Deep Learning.

Exploring Automatic Diagnosis of COVID-19 from Crowdsourced Respiratory Sound Data

Project Achoo: A Practical Model and Application for COVID-19 Detection from Recordings of Breath, Voice, and Cough

Wake-Cough: cough spotting and cougher identification for personalised long-term cough monitoring

Automatic cough detection from realistic audio recordings using C-BiLSTM with boundary regression