Respiratory signal analysis using PCA, FFT and ARTFA

Abstract

Sinus patients, both humans and animals, are increasing day by day in the world. That's why today signal analysis has been the need to know the diseases in the patient. Biomedical signal processing (BSP) has great importance in the life of every human and animal. Without BSP signals cannot be analysed, resulting in failure of disease acknowledgment. In this paper respiratory signals of Sinus and Normal Person has been analysed using Principal Component Analysis (PCA), Fast Fourier Transform (FFT) and Auto-Regressive Time-Frequency Analysis (ARTFA). PCA is used where dimension reduction is required. It has found many applications in BSP. ARTFA allows us to follow the changes in frequencies involved in the signal through time. For this, frequency changes in time are required to be observed. FFT examines the signal in frequency domain and calculates the spectral function (SF). In this paper, the variance of First Principal Component and Second Principal Component have been calculated for Sinus and Normal Person and these values are 86.94%, 13.05% and 92.733%, 7.266% respectively.