README.md documentation

README.md

@@ -1 +1,31 @@
 # Regularized K-SVD Algorithm
+
+Implementation of the [Regularized K-SVD](http://ieeexplore.ieee.org/abstract/document/7831399/) 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](http://www.cs.technion.ac.il/~ronrubin/Software/ompbox10.zip) 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](DL.m)).
+
+Have a look at the [test script](test_ksvd_reg.m) for a full example.