How can I prepare my dataset to fed into a stacked Autoencoder

With reference to this link https://in.mathworks.com/help/deeplearning/ug/train-stacked-autoencoders-for-image-classification.html

I am trying to implement stacked autoencoder for image classification. But I am not able to understand how can I prepare my dataset to fed into a autoencoder. As it is being said in this link that we need to reshape the training images into a matrix, how can it be done? Please provide a sample code.

Answer 1

[Ranjeet]

Hi Debojit,

The guidance on how to prepare dataset to fed into a stacked network has been provided in the following example –

Train Stacked Autoencoder for Image Classification

However, I am rewriting the sample code that serves the purpose –

% Get the number of pixels in each image 
imageWidth = 28; 
imageHeight = 28; 
inputSize = imageWidth*imageHeight; 

% Load the test images 
[xTestImages,tTest] = digitTestCellArrayData; 

% Turn the test images into vectors and put them in a matrix 
xTest = zeros(inputSize,numel(xTestImages)); 
for i = 1:numel(xTestImages) 
    xTest(:,i) = xTestImages{i}(:); 
end 
whos xTest xTestImages;

  Name               Size                 Bytes  Class     Attributes

  xTest            784x5000            31360000  double              
  xTestImages        1x5000            31880000  cell                

size(xTestImages{1})

ans = 1×2

28 28

You may find the code snippet in the example as well. The second last line in the code converts an image ‘xTestImages{i}’ into a vector and store in a matrix ‘xTest’.