Kernel-Based Multiview Joint Sparse Coding for Image Annotation

Algorithm 1

Optimizing method for problem (8).

Input:

  Training image samples ,

  Corresponding dictionary ,

  Regularization parameter and step size .

Output:

  Sparse coefficient matrix ,

(1) For  

(2) Initialization: Initialize , to be zero matrix; ; and set .

(3) While not converged Do

    Generalized gradient mapping step:

(4)   Calculate by Eq. (12)

(5)   For

(6)    Calculate th row of by Eq. (15)

(7)   End For

    Aggregation step:

(8)    

(9)     

(10)    

(11) End While

(12) End For