Adding some examples and cleaning up the code as well as switching to…

… md format for the readmes

README.md

@@ -0,0 +1,43 @@
+DeformIt - Tool for simulating ground truth labels and transformations via physical and statistical warps.
+==========================================================================================================
+Preet S. Jassi, May 5 2011
+--------------------------
+
+DeformIt is a Matlab tool to generate a large data-set deformed images and ground truth segmentations for machine learning algorithms based on a single prototype image and ground truth segmentation.  To deform 2D images, use 2D/DeformIt.m.  To deform 3D images use 3D/DeformIt.m
+
+To generate a large data-set of 2D images based on a single image and segmentation, run the following command in Matlab:
+
+```Matlab
+DeformIt(file, segmentation, nObs, WarpedImage, cp, RandomMorph, RandomRange, alpha1, t, Noise, NoiseType, NM,NV,ND,NonUniform, ncp, NonUniformStrength)
+```
+
+where:
+
+* _file_  - the filename
+* _segmentation_ - the segmentation name or empty string
+* _nObs_ - the number of images you want to generate
+* _WarpedImage_ - 1 if the image is to be deformed and 0 if the image is not to be deformed (warped)
+* _cp_ - the number of control points per axis of the image if you do not want to manually specify the control points
+* _RandomMorph_ - 1 if you want the deformation to be Random instead of Vibrational / Variational
+* _RandomRange_ - integer between [0,1] for the range of the random vectors where 1 is half the distance between control points
+* _alpha1_ - weight of vibrational vs variational when calculating the displacement vectors in the range [0,1]
+* _t_ - number of variational modes used in describing the deformation or the fractional variance if less than one, value between [0, nObs^2]
+* _Noise_ - 1 if you want to add noise to the image
+* _NoiseType_ - gaussian, poisson, salt & pepper, speckle
+* _NM_ - the mean of the noise for guassian noise
+* _NV_ - the variance for guassian and speckle
+* _ND_ - density for salt and pepper
+* _NonUniform_ - 1 if you want non uniformity intensities added to simulate the imhomogenity found in magentic resonance images
+* _ncp_ - the number of control points for the non uniformity
+* _NonUniformStrength_ - the strength of the non uniformity [0,1]
+
+You can also specify landmarks in a text file landmarks.txt and the new location of those landmarks will be specified.
+
+Generating deformed 3D images is very similar.  Look at DeformIt.m in the 3D folder for information about the parameters.
+
+You can read the paper here for more information about [DeformIt](http://www.erudianart.com/research/papers/deformit.pdf)
+
+
+
+
+

examples/2d/2d.m

@@ -0,0 +1,34 @@
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+% @file 2d.m
+% @author Preet S. Jassi
+% @date August 15th 2011
+% @description Example 2d deformation
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+
+% set the parameters for DeformIt to be used
+% detailed description of the parameters and their ranges is in the 
+% DeformIt.m file in the /src/2d/ directory
+% NOTE: this file needs to be copied into the source directory to work
+
+file = 'O1451U10X.tif';
+segmentation = 'O1451U10X_12042006.tif';
+nObs = '20';
+WarpedImage = '1';
+cp = '6';
+RandomMorph = '0';
+RandomRange = '0';
+alpha1 = '0.5';
+t = '0.8';
+Noise = '1';
+NoiseType = 'gaussian';
+NM = '0';
+NV = '0.01';
+ND = '0';
+NonUniform = '1';
+ncp = '6';
+NonUniformStrength = '0.5';
+
+% deform the image and the ground thruth segmentation
+% note that these files need to be copied into the source folder in order for it to work
+DeformIt(file, segmentation, nObs, WarpedImage, cp, RandomMorph, RandomRange, alpha1, t, Noise, NoiseType, NM,NV,ND,NonUniform, ncp, NonUniformStrength);

examples/3d/3d.m

@@ -0,0 +1,31 @@
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+% @file 3d.m
+% @author Preet S. Jassi
+% @date August 15th 2011
+% @description an example 3d image deformation
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+% set the parameters for DeformIt to be used
+% detailed description of the parameters and their ranges is in the 
+% DeformIt.m file in the /src/3d/ directory
+% NOTE: this file needs to be copied into the source directory to work
+
+filetype = 'analyze';
+file = 'analyze';
+segmentation = '0';
+nObs = '10';
+WarpedImage = '1';
+cp = '6';
+RandomMorph = '1';
+RandomRange = '0.5';
+alpha1 = '0';
+t = '0';
+Ansio = '1';
+Noise = '1';
+NonUniform = '1';
+ncp = '6';
+NonUniformStrength = '0.5';
+
+% deform the image and the ground thruth segmentation
+% note that these files need to be copied into the source folder in order for it to work
+DeformIt(filetype, file, segmentation, nObs, WarpedImage, cp, RandomMorph, RandomRange, alpha1, t, Ansio, Noise, NonUniform, ncp, NonUniformStrength); 

examples/README.md

@@ -0,0 +1,4 @@
+DeformIt Examples
+=================
+
+To execute the examples, first copy over the example code to the respective source directory (/src/2d/ or /src/3d/) and then execute 2d.m or 3d.m in Matlab in the /src/2d/ or /src/3d/ directories.

src/2d/DeformImage.m

@@ -0,0 +1,38 @@
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+% @file GenerateDeformedAndSegmentedImages.m
+% @author Preet S. Jassi (sorta...)
+% @date July, 31th 2008
+% @returns interpolated image
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+
+function [DeformedIm, X, Y] = DeformImage(Im,Frx,Fry,Tox,Toy,type)
+%  [x y x y x y x y x y] 
+[XX,YY] = meshgrid(1:size(Im,2), 1:size(Im,1));
+
+if (Frx == Tox' & Fry == Toy')
+    DeformedIm = Im;
+    X = XX;
+    Y = YY;
+    return
+end
+
+
+Rows = Frx;
+Columns = Fry;
+XDeformations = Tox;
+YDeformations = Toy;
+
+% Fit spline for the X Deformation
+X = FitSpline(XX, Rows, Columns, XDeformations);
+
+% Fit spline for the Y Deformation
+Y = FitSpline(YY, Rows, Columns, YDeformations);
+
+%noise types
+
+
+
+
+% Resample Image.
+DeformedIm = InterpolateImage2(X, Y, Im, type, XX, YY); 
+% return