Comparison of Neural Network Methods for Data Preprocessing in Solving Problems of Analysis, Diagnosis, and Classification of Defects of Radio-Electronic Equipment

We analyze the issues of preprocessing the data obtained from a 3D X-ray microtomograph when solving the problems of analysis, diagnosis, and classification of defects of radio-electronic equipment. It is proposed to use neural network methods and regularizing splines when solving the formulated problem. A comparative analysis of the neural network and spline methods is carried out when solving the problem of recovery of heavily noisy signals. The efficiency of the proposed approach has been studied by numerical simulation methods and when processing real data.