How to handle labels using fileDatastore in a CNN

I have a collection of 50x1x12 mat files, that I need to upload into some datastore to subsequently pass into a convolutional neural network, how can I handle the labels of this files?, what datastore should I use?,

I have been trying with file datastore, the filedatastore does not handle the labels in the same way as the imagedatastore does getting them from the folder.

-What datastore can I use

-if I use filedatastore how can I handle the labels after getting them from the foler and in the moment of use net = trainNetwork( ? );

-how the inputlayer should initially set

I really appreacite the help I have been stuck with this for along time.

Answer 1

[Madhav Thakker]

Hi Mauricio,

You can use two different datastores for reading the .mat files and labels separately. Then, you can use combine on the datastores and feed to the CNN network.

Refer https://in.mathworks.com/matlabcentral/answers/459161-image-regression-using-mat-files-and-a-datastore#answer_405285 answer for a similar approach.

Hope this helps.

[Mauricio Piraquive]

Thanks for you answer,

I am doing this following the example

in the target Data I am using this function to get the labels from the folders

function label = readLabel(filename,classNames)
filepath = fileparts(filename);
[~,label] = fileparts(filepath);

label = categorical(string(label),classNames);
end

and then following the example I am uploading the files into the fileDatastore

inputData=fileDatastore(fullfile('inputData'),'ReadFcn',@load,'FileExtensions','.mat');

classNames = string(1:2)
targetData=fileDatastore(fullfile('targetData'),'ReadFcn',@load,'FileExtensions','.mat',readLabel(filename,classNames));

Should I trasnform before combine the filedatastores? if so, how should my transform function be?

I tried without transforming and i got this error:

"The training images are of size 1x1x1 but the input layer expects images of size 50x1x12"

Thanks for your help

[Saif Aljanahi]

Dear Mauricio, did you conducts this example and you got results? I really need your help in this matter.

I'm want to label such data from datastore or folder data!.

Thanks.

[Pierre]

Why use 2 different datastores and combine them when you can just 1 with both predictors and response?