When I add new image , I want to get "the result" of image in neural network pattern recognition. How can I do this according to codes added?

I am studying about mammogram images to detect cancer on breast image. There are three situations benign, malign and normal.

I trained my neural netwok with 100 images and get result 95% . I will create user interface and will select an image and get the result of what is this image benign or malign or normal .

I mean my training is done and I want to see my neural network ability to detect result. I will upload a new image in user interface and want to see result of image benign or malign or normal . But I don't know which codes can I add in neural network?

Here is code neural pattern recognition created by matlab ownself. Which code should be added to detect new image (for example for 101 th image) ?

% Solve a Pattern Recognition Problem with a Neural Network
% Script generated by Neural Pattern Recognition app
% Created 07-Feb-2021 15:50:44
%
% This script assumes these variables are defined:
%
%   x - input data.
%   y - target data.

x = x;
t = y;

% Choose a Training Function
% For a list of all training functions type: help nntrain
% 'trainlm' is usually fastest.
% 'trainbr' takes longer but may be better for challenging problems.
% 'trainscg' uses less memory. Suitable in low memory situations.
trainFcn = 'trainscg';  % Scaled conjugate gradient backpropagation.

% Create a Pattern Recognition Network
hiddenLayerSize = 10;
net = patternnet(hiddenLayerSize, trainFcn);

% Setup Division of Data for Training, Validation, Testing
net.divideParam.trainRatio = 70/100;
net.divideParam.valRatio = 15/100;
net.divideParam.testRatio = 15/100;

% Train the Network
[net,tr] = train(net,x,t);

% Test the Network
y = net(x);
e = gsubtract(t,y);
performance = perform(net,t,y)
tind = vec2ind(t);
yind = vec2ind(y);
percentErrors = sum(tind ~= yind)/numel(tind);

% View the Network
view(net)

% Plots
% Uncomment these lines to enable various plots.
%figure, plotperform(tr)
%figure, plottrainstate(tr)
%figure, ploterrhist(e)
%figure, plotconfusion(t,y)
%figure, plotroc(t,y)

Answer 1

[Abhishek Gupta]

Hi,

As per my understanding, you want to make predictions for new input using your trained network. You can do the same using the 'predict()' function in MATLAB: -

predictions = predict(trainedNetwork,newImages);

For more information, check out the documentation here: -

https://www.mathworks.com/help/deeplearning/ref/seriesnetwork.predict.html

[Ali Zulfikaroglu]

Actually I tried this way but it doesn't work . "a" is my data has glcm features. I also tried with only image and wrote " img" for my image instead of "a" and still didn't work.

predictions = predict(net,a);

%İt gave that error at the bottom.
Check for missing argument or incorrect argument data type in call to function 'predict'.

Error in testnetwork (line 106)
predictions = predict(net,a);

[Abhishek Gupta]

You need to make sure that the input argument data-type is correct. For more information related to input arguments, check out the link below: -

https://www.mathworks.com/help/deeplearning/ref/seriesnetwork.predict.html#d122e79317

[Ali Zulfikaroglu]

I used glcm feature extraction methods to detect properties of image and glcm gives 88 properties (such as homogenity, entropy, energy...). I wrote 88 numbers in excel file and these are my input and I write output according to this 88 properties . I have data in excel file . In excel file , there is 88x100 input and 2x100 target. The code above x= input and y= my target data. I trained my network with this inputs and targets . You can see code above. And then I added new image and get glcm properties of it again 88 features . And I wonder its output "what" according to this 88 features .

predictions = predict(net,a);

net is my trained network and a is that new image's 88 properties.

and this code doen't work.

But I used new_outputs=sim(net,a) it worked.

Should we use "sim" command or "predict" command ? "sim" command is true in here?

[Abhishek Gupta]

What error are you getting while using "predict()"? What is the dimension of 'a'? Note that "a" should be an N-by-numFeatures numeric array, where N is the number of observations and numFeatures is the number of features of the input data. I see N=1 in your case, so "a" should be of (1x88) dimension.

For more information, check out the "Input Arguments" section of the documentation: -

https://www.mathworks.com/help/deeplearning/ref/seriesnetwork.predict.html#mw_0a51db93-cccf-4b2f-ae4c-6724cbf5ec46_sep_shared-X