Preparing a datastore for a multi-input CNN

I have written the following code based on a question and answer on mathworks website. The problem is that the resulting datastore has only one row:

root_train='E:\DeepNet\Combined\tr';

trdatstore = imageDatastore(root_train,'IncludeSubfolders',true,'LabelSource','foldernames');

imds = trdatstore;

augimdsTrain = augmentedImageDatastore([224 224 3],imds)

augimdsTrain.MiniBatchSize =1

taugimdsTrain = transform(augimdsTrain,@(x)x{1,1})

labelsTrain = transform(augimdsTrain,@(x){x{1,2}})

cds = combine(taugimdsTrain,taugimdsTrain,labelsTrain)

Answer 1

[yanqi liu]

visiondataPath = fullfile(matlabroot,'toolbox','vision','visiondata');
trdatstore = imageDatastore(visiondataPath,'IncludeSubfolders',true,'LabelSource','foldernames');
imds = trdatstore; 
augimdsTrain = augmentedImageDatastore([224 224 3],imds)

augimdsTrain = 
  augmentedImageDatastore with properties:

             NumObservations: 2450
                       Files: {2450×1 cell}
    AlternateFileSystemRoots: {}
               MiniBatchSize: 128
            DataAugmentation: 'none'
          ColorPreprocessing: 'none'
                  OutputSize: [224 224]
              OutputSizeMode: 'resize'
        DispatchInBackground: 0

augimdsTrain.MiniBatchSize =1

augimdsTrain = 
  augmentedImageDatastore with properties:

             NumObservations: 2450
                       Files: {2450×1 cell}
    AlternateFileSystemRoots: {}
               MiniBatchSize: 1
            DataAugmentation: 'none'
          ColorPreprocessing: 'none'
                  OutputSize: [224 224]
              OutputSizeMode: 'resize'
        DispatchInBackground: 0

taugimdsTrain = transform(augimdsTrain,@(x)x{1,1})

taugimdsTrain = 
  TransformedDatastore with properties:

      UnderlyingDatastores: {augmentedImageDatastore}
    SupportedOutputFormats: ["txt"    "csv"    "xlsx"    "xls"    "parquet"    "parq"    "png"    "jpg"    "jpeg"    "tif"    "tiff"    "wav"    "flac"    "ogg"    "mp4"    "m4a"]
                Transforms: {@(x)x{1,1}}
               IncludeInfo: 0

labelsTrain = transform(augimdsTrain,@(x){x{1,2}})

labelsTrain = 
  TransformedDatastore with properties:

      UnderlyingDatastores: {augmentedImageDatastore}
    SupportedOutputFormats: ["txt"    "csv"    "xlsx"    "xls"    "parquet"    "parq"    "png"    "jpg"    "jpeg"    "tif"    "tiff"    "wav"    "flac"    "ogg"    "mp4"    "m4a"]
                Transforms: {@(x){x{1,2}}}
               IncludeInfo: 0

cds = combine(taugimdsTrain,taugimdsTrain,labelsTrain)

cds = 
  CombinedDatastore with properties:

      UnderlyingDatastores: {[1×1 matlab.io.datastore.TransformedDatastore]  [1×1 matlab.io.datastore.TransformedDatastore]  [1×1 matlab.io.datastore.TransformedDatastore]}
    SupportedOutputFormats: ["txt"    "csv"    "xlsx"    "xls"    "parquet"    "parq"    "png"    "jpg"    "jpeg"    "tif"    "tiff"    "wav"    "flac"    "ogg"    "mp4"    "m4a"]

[mingsheng wang]

你好，这个多输入的网络智能用datastore吗？能不能直接用table呢？