Development of Deep Convolutional Neural Network with Adaptive Batch Normalization Algorithm for Bearing Fault Diagnosis
Table 2
Adaptive WFK-CNN algorithm based on Ad-BN domain.
Adaptive WFK-CNN algorithm based on Ad-BN domain
Input
Target domain signal p, expression of the i neuron in the BN layer of WFK-CNN , among them , for the i neuron, trained zoom and shift parameters and
Output
Adjusted WFK-CNN network
For
Each neuron i and each signal p in the target domain Calculate the mean and variance of all samples in the target domain: Calculate the output of BN layer: