A Microphone Calibration Scheme for Cough Sound Analysis and Diagnosis

A self-calibration system implemented as part of an audio signal processing mobile app with dithering and digital filtering is presented for cough sound analysis and subsequent remote diagnosis of the type of cough cause, the cough itself can be due to different conditions such as chemical irritation, emphysema, pneumonia, bacterial or viral infection. The algorithm will produce a signal that normalizes the input cough sound irrespective of the type of smartphone used, the computed information makes analysis possible with audio signal collected from microphones produced by different manufacturers that may exhibit different sonic properties.

I. Introduction

Cough is a frequent medical condition due to numerous possible causes, ranging from simple short-term throat irritation, bacterial or viral infection, to COPD (chronic obstructive pulmonary disease). Clinical visits in rural areas or by patients with impaired mobility for diagnosis can be simplified by utilizing telemedicine for remote diagnosis and automatic electronic patient record (ePR) updating [1]. Diagnosis in cough using a smartphone based platform and recommend subsequent treatment without clinic visit requires the acquisition of cough sound suitable for audio analysis with devices from any manufacturer of varying quality. Remote diagnosis entails mapping the received audio signal of a patient's cough to a reference pattern that contains information such as circadian distribution and frequency representation so that the cause of cough can be determined.