How to know the data set after augment?

Hello everyone,

I was developed the coding for augmented my data for trainning process using 3D U-Net network. But I dont know how to find the addition data after augment? Because I checked in workspace but not found. Below is my coding, the function @augment3dPatch is for augment data

Also I attached the function for running my coding.

The data for image training (ImagesTr), groundtruthlabeling training (LabelsTr), image validation (imagesVal ), groundtruth validation (labelsVal) can download thru this link https://drive.google.com/drive/folders/176Ffn3d-61HFYiK7ywFFD9rRuzmXnN26?usp=share_link

clc
clear all
close all
%testDataimages
DATASetDir = fullfile('C:\Users\Akmal\Desktop\NEW 3D U NET');
IMAGEDir = fullfile(DATASetDir,'ImagesTr');
volReader = @(x) matRead(x);
volds = imageDatastore(IMAGEDir, ...
    'FileExtensions','.mat','ReadFcn', volReader);
% labelReader = @(x) matread(x);
matFileDir = fullfile('C:\Users\Akmal\Desktop\NEW 3D U NET\LabelsTr');
classNames = ["background", "tumor"];
pixelLabelID = [0 1];
% pxds = (LabelDirr,classNames,pixelLabelID, ...
%     'FileExtensions','.mat','ReadFcn',labelReader);
pxds = pixelLabelDatastore(matFileDir,classNames,pixelLabelID, ...
    'FileExtensions','.mat','ReadFcn',@matRead);
volume = preview(volds);
label = preview(pxds);
up1 = uipanel;
h = labelvolshow(label, volume, 'Parent', up1);
h.CameraPosition = [4 2 -3.5];
h.LabelVisibility(1) = 0;
h.VolumeThreshold = 0.5;
volumeViewer(volume, label)
patchSize = [128 128 64];
patchPerImage = 16;
miniBatchSize = 8;

patchds = randomPatchExtractionDatastore(volds,pxds,patchSize, ...
    'PatchesPerImage',patchPerImage);
patchds.MiniBatchSize = miniBatchSize;

%% AUGMENT DATA FOR TRAINING
dsTrain = transform(patchds,@augment3dPatch);

volLocVal = fullfile('C:\Users\Akmal\Desktop\NEW 3D U NET\imagesVal');
voldsVal = imageDatastore(volLocVal, ...
    'FileExtensions','.mat','ReadFcn',volReader);
lblLocVal = fullfile('C:\Users\Akmal\Desktop\NEW 3D U NET\labelsVal');
pxdsVal = pixelLabelDatastore(lblLocVal,classNames,pixelLabelID, ...
    'FileExtensions','.mat','ReadFcn',volReader);
dsVal = randomPatchExtractionDatastore(voldsVal,pxdsVal,patchSize, ...
    'PatchesPerImage',patchPerImage);
dsVal.MiniBatchSize = miniBatchSize;

%% AUGMENT DATA FOR VALIDATION
 dsVal = transform(patchds,@augment3dPatch); 
 
inputSize = [128 128 64];
numClasses = 2;
encoderDepth = 2;
lgraph = unet3dLayers(inputSize,numClasses,'EncoderDepth',encoderDepth,'NumFirstEncoderFilters',16) 
figure,plot(lgraph);

Answer 1

[Walter Roberson]

But I dont know how to find the addition data after augment? Because I checked in workspace but not found.

Augmented data is generated internally and is not stored after it is used.If you need it then you can use readall()

Answer 2

[Askic V]

This is an extract from the documentation:

% transform   Create a new datastore that applies a function to the 
%     data read from all input datastores.
%  
%     DSNEW = transform(DS, transformFcn) transforms an input
%     Datastore, DS, given an input function, transformFcn. The
%     transform function has the syntax:
%  
%         dataOut = FCN(dataIn);
%  
%     where FCN is a function that takes the dataIn returned by
%     read of DS and returns a modified form of that data,
%     dataOut.

In your case the new datastorage after augmenting will be dsTrain

How you can inspect the newdata depends on what transform function do.

For example, this is a simple case of image resizing as transformFcn

imds = imageDatastore('peppers.png');
img1 = read(imds); % reads theimage
subplot(211)
imshow(img1)
newDS = transform(imds,@(x) imresize(x, [120, 120]));
imgOut = read(newDS);
subplot(212)
imshow(imgOut);

whos

  Name          Size                Bytes  Class                                       Attributes

  cmdout        1x33                   66  char                                                  
  imds          1x1                     8  matlab.io.datastore.ImageDatastore                    
  img1        384x512x3            589824  uint8                                                 
  imgOut      120x120x3             43200  uint8                                                 
  newDS         1x1                     8  matlab.io.datastore.TransformedDatastore              

As you can see, I need to know how to read the data behind, because datastore is just a pointer to location where the data is.

I hope this example will help you figuring out how exactly to inspect your data.

[mohd akmal masud]

I see..you means is I need to read the dsTrain?

[mohd akmal masud]

as you can see in attched article, page 7, the augmented data become 2280 from the original data set 112.

Thats I need to know in my coding.