I. Introduction The MATLAB code contained in this directory implements the algorithm derived in the paper [1] using Majorization-Minimization technique. This file describes the details and procedures for the examples contained in [1].

II. Files contained and their description This directory contains the following files-

 1. demo.m
     This file demonstrates the CNC FLSA [1] for denoising a 
     sparse piecewise constant signal. This file reproduces example 1 
     in the paper [1]. 
     
 3. ECG_demo.m
     This file demonstrates the denoising of a synthetic ECG signal, 
     generated using ECGSYN (see [1] for details). 
     The synthetic ECG signal is obtained by the following commands
     
     fs = 256;
     ecg = ecgsyn(fs, 20); % 20 is the number of beats to be simulated.                                                   
  
     
 5. CNC_FLSA.m
     This function minimizes the CNC FLSA objective function
     F(x) = 0.5||y-x||_2^2 + lam0*phi(x,a0) + lam1*phi(Dx,a1)
     using the majorization-minimization technique, where phi
     is a non-convex penalty function. type `help CNC_FLSA' for more details
      
 6. soft.m
     This file implements the soft thresholding rule. 
     type `help soft.m' for more details
 
 7. tvd.c, tvd.mexmaci, tvd.mexmaci64, and tvd.mex64
     C++ implementation of TV denoising (see [1] for details)

For questions/comments contact: Ankit Parekh ([email protected])

Please cite as: [1] Convex fused lasso denoising with non-convex regularization and its use for pulse detection. Ankit Parekh and Ivan W. Selesnick, IEEE SPMB, 2015.