A Variational Bayesian Superresolution Approach Using Adaptive Image Prior Model

Algorithm 1

Our iterative SR algorithm.

Input:

  The observations ; The motion parameters ; The hyper-parameters ,

  ; The scale parameters ; The parameters .

Output:

  The HR image ;

(1) Use the Bicubic interpolation of the first LR image as initial value for the HR image.

(2) ;

(3) repeat

(4)  Compute by using (20) and (21);

(5)  Compute by using (22);

(6)  Compute by using (23) and (24);

(7)  Compute by using (25);

(8)  Compute by using (26) and (27);

(9)  Compute by using (14);

(10)  ;

(11) until convergence criterion is met

(12) Set .