Combination of Wavelets and Hard Thresholding for Analysis of Cough Signals

Cough signals are fundamental symptom of respiratory diseases. During the acquisition of the cough signals via microphone, it gets contaminated due to the noise present in the surroundings or certain other reasons. It is a tough task to remove this noise from contaminated cough sound signals. This call for mathematical analysis of cough signals to aid in diagnosis of respiratory diseases. This paper consists of combination of hard thresholding and Discrete Wavelet Transform (DWT) for eliminating the noise present in the acquired cough sound signals. An improved result of noise filtering has been demonstrated in the result analysis with ‘Sym4’ wavelet family. This has been supported with better values of Signal-to-Noise Ratio (SNR) of the acquired cough signals.