CoughNet-V2: A Scalable Multimodal DNN Framework for Point-of-Care Edge Devices to Detect Symptomatic COVID-19 Cough

With the emergence of COVID-19 pandemic, new attention has been given to different acoustic bio-markers of the respiratory disorders. Deep Neural Network (DNN) has become very popular with the audio classification task due to its impressive performance for speech detection, audio event classification etc. This paper presents CoughNet-V2 - a scalable multimodal DNN framework to detect symptomatic COVID-19 cough. The framework was designed to be implemented on point-of-care edge devices to help the doctors at pre-screening stage for COVID-19 detection. A crowd-sourced multimodal data resource which contains subjects' cough audio along with other relevant medical information was used to design the CoughNet-V2 framework. CoughNet-V2 shows multimodal integration of cough audio along with medical records improves the classification performance than that of any unimodal frameworks. Proposed CoughNet-V2 achieved an area-under-curve (AUC) of 88.9% for the binary classification task of symptomatic COVID-19 cough detection. Finally, measurement of the deployment attributes of the CoughNet-V2 model onto processing components of an NVIDIA TX2 development board is presented as a proposition to bring the healthcare system to consumers' fingertips. Clinical relevance - CoughNet-V2 will help medical practitioners to asses whether the patients need intensive medical help without physically interacting with them. © 2022 IEEE.

CoughNet: A Flexible Low Power CNN-LSTM Processor for Cough Sound Detection

The continuing effect of COVID-19 pulmonary infection has highlighted the importance of machine-aided diagnosis for its initial symptoms such as fever, dry cough, fatigue, and dyspnea. This paper attempts to address the respiratory-related symptoms, using a low power scalable software and hardware framework. We propose CoughNet, a flexible low power CNN-LSTM processor that can take audio recordings as input to detect cough sounds in audio recordings. We analyze the three different publicly available datasets and use those as part of our evaluation to detect cough sound in audio recordings. We perform windowing and hyperparameter optimization on the software side with regard to fitting the network architecture to the hardware system. A scalable hardware prototype is designed to handle different numbers of processing engines and flexible bitwidth using Verilog HDL on Xilinx Kintex-7 160t FPGA. The proposed implementation of hardware has a low power consumption of o 290 mW and energy consumption of 2 mJ which is about 99 x less compared to the state-of-the-art implementation.