A few Functions for manipulating images

6 A few Functions for manipulating images

This functions are not well documented yet

6.1 Function mexExtractPatches

%
% Usage: X =mexExtractPatches(I,n,step);
%
% Name: mexExtractPatches
%
% Description: Extract n x n patches spaced every ``step'' pixels from
% an image I of size nx x ny x nchannels
%
% Inputs: I: double nx x ny x nchannels
%
% Output: X: double dense matrix
%
% Author: Julien Mairal, 2014

6.2 Function mexCombinePatches

%
% Usage:   I =mexCombinePatches(X,I0,n,step,lambda,normalize);
%
% Name: mexCombinePatches
%
% Description: Combine patches extracted with mexExtractPatches into a new image
%              I = lambda I0 + combined(X) if normalize, averaging is
%              performed; otherwise the patches are just summed.
%
% Inputs: I0:  double nx x ny x nchannels
%
% Output: I: double nx x ny x nchannels
%
% Author: Julien Mairal, 2014

6.3 Function mexConvFistaFlat

%
% Usage:   alpha =mexConvFistaFlat(I,D,alpha0,param);
%
% Name: mexConvFistaFlat
%
% Description: performs convolutional sparse encoding of an
%              image I with a local dictionary D, using FISTA
%              and similar options as mexFistaFlat
%
% Inputs: I:  double nx x ny x nchannels
%         D:  dictionary
%         alpha0: initial weights
%
% Output: alpha: output coefficients
%
% Author: Julien Mairal, 2014