Title :
Life detection based on Correntropy Spectral Density
Author :
Kong, Lingjiang ; Principe, Jose C.
Author_Institution :
Dept. Electron. Eng., Univ. of Electron. Sci. & Technol. of China, Chengdu, China
Abstract :
A life detection method based on Correntropy Spectral Density (CSD) is proposed. Because correntropy contains second and higher-order moments of the probability density function due to the nonlinearty of the kernel, it is capable of yielding higher output Signal to Noise Ratio (SNR) in non-Gaussian background when compared with traditional life detection algorithm based on the Power Spectral Density (PSD). The simulation with synthetic data and experimental results illustrate the superiority of the method.
Keywords :
probability; signal detection; spectral analysis; CSD; PSD; SNR; correntropy spectral density; higher-order moments; life detection algorithm; life detection method; nonGaussian background; power spectral density; probability density function; signal to noise ratio; synthetic data; Fourier transforms; Heart beat; Kernel; Loudspeakers; Radar; Signal to noise ratio; Stochastic processes; correntropy spectral density (CSD); life detection; non-Gaussian;
Conference_Titel :
Signal Processing (ICSP), 2010 IEEE 10th International Conference on
Conference_Location :
Beijing
Print_ISBN :
978-1-4244-5897-4
DOI :
10.1109/ICOSP.2010.5657116
