The Deep Learning code for paper Metallic Plate Permeability Estimation using Single Frequency Eddy Current Testing in the Presence of Probe Lift-off
The architecture of the DL model is the modified ResNet18-1D. The inductance spectrum is fed to the input layer of the model, while other input variables are appended to the internal layer of the model, cascading with low-dimensional feature vectors extracted from the inductance spectrum.
Get dataset from Kaggle (https://www.kaggle.com/datasets/rusuanjun/plate-permeability-dataset). Run the train.py to train the network.