Skip to content

Latest commit

 

History

History
31 lines (22 loc) · 1004 Bytes

README.md

File metadata and controls

31 lines (22 loc) · 1004 Bytes
Raw

Regularized K-SVD Algorithm

Implementation of the Regularized K-SVD Dictionary Learning Algorithm described in B. Dumitrescu and P. Irofti, "Regularized K-SVD," in IEEE Signal Processing Letters, vol. 24, no. 3, pp. 309-313, March 2017.

Prerequisite

OMP implementation by Ron Rubinstein

Usage

INPUTS:

  • Y -- training signals set
  • D -- current dictionary
  • X -- sparse representations
  • iter -- current DL iteration

PARAMETERS:

  • reg -- regularization factor (default: 0.01)
  • vanish -- regularization vanishing factor (default: 0.95)
  • regstop -- stop regularization from this iteration on (default: Inf)

OUTPUTS:

  • D -- updated dictionary
  • X -- updated representations

Sample call

[D,X] = ksvd_reg(Y,D,X,iter)

to be used within a dictionary learning loop (see DL).

Have a look at the test script for a full example.