A Multi-Branch Deep Learning Network for Automated Detection of COVID-19

Abstract

Fast and affordable solutions for COVID-19 testing are necessary to contain the spread of the global pandemic and help relieve the burden on medical facilities. Currently, limited testing locations and expensive equipment pose difficulties for individuals trying to be tested, especially in low-resource settings. Researchers have successfully presented models for detecting COVID-19 infection status using audio samples recorded in clinical settings [5, 15], suggesting that audio-based Artificial Intelligence models can be used to identify COVID-19. Such models have the potential to be deployed on smartphones for fast, widespread, and low-resource testing. However, while previous studies have trained models on cleaned audio samples collected mainly from clinical settings, audio samples collected from average smartphones may yield suboptimal quality data that is different from the clean data that models were trained on. This discrepancy may add a bias that affects COVID-19 status predictions. To tackle this issue, we propose a multi-branch deep learning network that is trained and tested on crowdsourced data where most of the data has not been manually processed and cleaned. Furthermore, the model achieves state-of-art results for the COUGHVID dataset [16]. After breaking down results for each category, we have shown an AUC of 0.99 for audio samples with COVID-19 positive labels.

CovidEnvelope: An Automated Fast Approach to Diagnose COVID-19 from Cough Signals

“…Thus, we selected threshold 3,000 for further analyses. To compare data processing and inference times across different models, we used verified and matched contidions of Russian dataset in our CovidEnvelope approach and two other state of art methods [18], [20]. Our approach is significantly faster in data preprocessing and inference time which is shown in Figure 4.…”

“…To compare among different models, we compared our CovidEnvelope approach with the state of art models [18], [20]. For COVID-19 cough detection, we used Russian dataset, including different conditions of "Verbal", "Verified", "Matched", "MatchedAsymp" and "MatchedSymp".…”

“…Laguarta et al [19] considered CNN for diagnosing COVID-19 cough from extracted features. Moreover, Fakhry et.al [20] used deep neural networks to diagnose COVID-19 by considering clinic records and mel-frequency cepstral coefficients and melfrequency spectrograms of cough audios. However, they [18], [20] used "verbal" confirmed cough sounds, which can lead to false-positive or negatives in COVID-19 detection.…”

“…Moreover, Fakhry et.al [20] used deep neural networks to diagnose COVID-19 by considering clinic records and mel-frequency cepstral coefficients and melfrequency spectrograms of cough audios. However, they [18], [20] used "verbal" confirmed cough sounds, which can lead to false-positive or negatives in COVID-19 detection. Both of these models are computationally expensive, require heavy data prepossessing and inference time.…”

“…Finding a reliable and well-balanced dataset is pre-requisite to build and test ML model, but there is a scarcity of COVID-19 cough audio dataset. Most of the existing models [20], [21] used Coughvid, Cowsara, Sarcos and Virufy dataset, which lack proper annotation and corresponding confirmation. Coughvid dataset mostly contains cough audios from verbal confirmation, and for PCR verified audios, it only contains COVID-19 positive samples.…”