Application of Wavelet Packet Decomposition and Deep Belief Network for Rectal Function Diagnosis

Anal incontinence refers to as the loss of the ability of the body to store and control the liquid and solid contents and gases in the rectum, seriously affecting the normal life of patients. To accurately understand the health status of rectal function of the human body, a rapid diagnostic model for rectal function was proposed in this paper. Aiming at overcoming the deficiency of signal feature extraction, this paper used wavelet packet decomposition to extract the rectal pressure signals that was obtained in a real-time manner. Through the training and testing of the extracted feature parameters, the diagnosis results were reported in the form of classification. The experimental results showed that wavelet packet decomposition could well separate the characteristic parameters of rectal signal. The diagnosis model developed in this study can achieve the diagnosis of rectal function, with an average diagnosis rate of 92.1008%.