Abstracto

Recognition of congenital heart sound signals based on wavelet denoising and bispectrum

Chen Tian-hua, Xiang Ling-zi

Acquiring the heart sound signal without noise is the premise and foundation to realize noninvasive diagnosis of congenital heart disease. But the heart sound signal belongs to the weak biological signals of the human body on the strong noise background. Acquiring the signal is susceptible to noise interference, so how to eliminate the inside noise of heart sounds is the key to distinguish the childrenÂ’s normal and abnormal heart sounds. The heart sound signals acquisition system based on DSP is proposed in this paper, using db6wavelet filter to smoothing the collected 21 cases of childrenÂ’s heart sound signals, which achieved the very good de-noising effect. Estimating these signals with bi-spectrumcan accurately identify 3 cases of congenital heart disease in which the heart sound signals. Experiments showthat this method can better identify the normal heart sounds and children congenital heart sound signals.