Classifier Training: Eight-layer CNNs trains different levels of rotation and deformation images
Input: Images , and initial velocity field
Output:Transformation from to
Pre-registration: For 2D image, through CNNs classifier to rectify rotation then scale and transformation;
For 3D image, triplanar 2D CNNs to pre-registrate
Demons registration:
 (i) Find updates mapping to
 (ii) Smooth updates:
 (iii) Update velocity field:
      (approximated with )
 (iv) Smooth velocity field:
 (v) Use PCA to extract features and calculate new similarity metric as PCA-SSD, PCA-pearson,
    PCA-spearman, and kendall.
until convergence
Algorithm 3: Two-layer unsupervised deep adaptive registration framework.