In recent time, COVID-19 pandemic has a devastating impact on our health and social well-being. Besides developing vaccines, numerous approaches has been utilised for COVID-19 detection. Cough audio signals classification showed potential to serve as a pre-screening tool and numerous costly and sophisticated deep learning algorithms are being implemented with good accuracy. We developed a low-cost envelope approach, called CovidEnvelope, which can classify COVID-19 positive and negative cases from raw cough audio data and doesn’t need sophisticated computing resources. Due to availability of more reliable datasets, now there is much scope to increase the accuracy and implement in HCI healthcare and wellbeing sectors.
- Train current model with reliable COVID-19 positive and negative cough dataset to increase accuracy.
- Train and validate current “CovidEnvelope” algorithm with cough sounds, resulting from numerous respiratory tract pathologies.
- Develop mobile based application for public health.
1. Hossain, M.Z., Uddin, M.B. and Ahmed, K.A., 2021. CovidEnvelope: A Fast Automated Approach to Diagnose COVID-19 from Cough Signals. medRxiv.