PCANet is an unsupervised Convolutional Neural Network (CNN), which uses Principal Component Analysis (PCA) to learn convolutional filters. One drawback of that linear PCA cannot capture nonlinear structures within data. To address this problem, a straightforward approach utilizing kernel methods by equipping the method in with function. However, practice leads network having cubic complexity r...