2019 Chinese Control Conference (CCC) | 2019
A Diagnostic Model of Rectal function Reconstruction Based on Wavelet Analysis and LVQ Neural Network
Abstract
In view of the problems of the existing rectal pressure measurement and diagnosis, we recommend the use of an artificial anal sphincter system for pressure collection and diagnosis of the intestine. In this paper, the semi-soft threshold denoising method of wavelet transform is used to denoise the pressure signal collected by the artificial anal sphincter actuator with sensors, and the denoised noise is decomposed by six-layer wavelet transform. Then, we extract the high and low frequencies, reconstructed high and low frequencies characteristic components. Quantitative feature is the input of neural network, normal and abnormal intestinal types are the output of network. The Leaming Vector Quantization (LVQ) neural network is trained with training set data, and then the testing data are tested. Finally, the test results are analyzed. The results of simulation test show that the device can help doctors identify the intestinal function of patients and provide a basis for further diagnosis.