Adapting 1D CNN

Hello, I'm trying to adapt a 1D-CNN, which is originally described in https://arxiv.org/abs/1610.01683, to my own samples, which have the following format:

4 classes of signals, 30 samples per class, each sample is a one-dimensional array with 100 points.

From a previous post (https://www.mathworks.com/matlabcentral/answers/331164-convolutional-1d-net) I found the following architecture for the network:

inputLayer=imageInputLayer([1 6000]);
c1=convolution2dLayer([1 200],20,'stride',1);
p1=maxPooling2dLayer([1 20],'stride',10);
c2=convolution2dLayer([20 30],400,'numChannels',20);
p2=maxPooling2dLayer([1 10],'stride',[1 2]);
f1=fullyConnectedLayer(500);
f2=fullyConnectedLayer(500);
s1=softmaxLayer;
outputLayer=classificationLayer;

convnet=[inputLayer; c1; p1; c2; p2; f1; f2; s1;outputLayer]

opts = trainingOptions('sgdm');
convnet = trainNetwork(allData',labels,convnet,opts);

Output:

convnet = 

    9x1 Layer array with layers:

       1   ''   Image Input             1x6000x1 images with 'zerocenter' normalization
       2   ''   Convolution             20 1x200 convolutions with stride [1  1] and padding [0  0]
       3   ''   Max Pooling             1x20 max pooling with stride [10  10] and padding [0  0]
       4   ''   Convolution             400 20x30 convolutions with stride [1  1] and padding [0  0]
       5   ''   Max Pooling             1x10 max pooling with stride [1  2] and padding [0  0]
       6   ''   Fully Connected         500 fully connected layer
       7   ''   Fully Connected         500 fully connected layer
       8   ''   Softmax                 softmax
       9   ''   Classification Output   cross-entropy

I changed some of the parameters to adapt the networks for my samples, namely:

inputLayer=imageInputLayer([1 100]);    % [1 6000] replaced by [1 100]
c1=convolution2dLayer([1 20],20,'stride',1);   % [1 200] replaced by [1 20]
p1=maxPooling2dLayer([1 20],'stride',10);
c2=convolution2dLayer([20 30],400,'numChannels',20);
p2=maxPooling2dLayer([1 10],'stride',[1 2]);
f1=fullyConnectedLayer(500);
f2=fullyConnectedLayer(500);
s1=softmaxLayer;
outputLayer=classificationLayer;

I tried to run this same network but I got an error regarding the dimensions of Layer 4:

Layer 4: Input size mismatch. Size of input to this layer is different from the expected
    input size.
    Inputs to this layer:
        from layer 3 (output size 1×7×20)

Could you help me finding the appropiate dimensions for network's architecture? Thanks...

Answer 1

[Mahesh Taparia]

Hi

Try to change the maxpooling operation,like make it with small window like [1 2] or you can remove max pooling operation as the input dimension is not that large. The error you are getting because of size mismatch between features and the hidden layer parameters.

Hope it will help!

[Fernando Meneses]

Thank you! Removing the maxpooling layer worked.