App designer error classify

Dear all,

while creating an app using app designer I am not able to compeltelly deploy my trainned CNN. My main issue is wiht the function classify which should be valid with:

[YPred]= classify(app.net,app.imds1);

Where app.net is the trainned CNN, app.imds1 is the imageDatastore where all the images that I want to classify are stored.

The error message that I got is:

Error using classify (line 123)

Requires at least three arguments.

Meaning that the function classify is using a 2nd options instead of the one that I want it to use. Is there a way to make it work? Could it be that I did not completelly unistall Matlab 2018b from my computer? I have 2019a running but it never occurred this error before...

The full code is the following one:

classdef ICapp < matlab.apps.AppBase

% Properties that correspond to app components

properties (Access = public)

UIFigure matlab.ui.Figure

DistanceEvolution matlab.ui.control.UIAxes

UITable matlab.ui.control.Table

Load matlab.ui.control.StateButton

Predict matlab.ui.control.StateButton

FrameNSpinnerLabel matlab.ui.control.Label

FrameNSpinner matlab.ui.control.Spinner

Save matlab.ui.control.Button

DistanceextrantionappLabel matlab.ui.control.Label

Image matlab.ui.control.Image

Image_2 matlab.ui.control.Image

PLOT matlab.ui.control.UIAxes

DoneLampLabel matlab.ui.control.Label

DoneLamp matlab.ui.control.Lamp

DoneLamp_2Label matlab.ui.control.Label

DoneLamp_2 matlab.ui.control.Lamp

DoneLamp_3Label matlab.ui.control.Label

DoneLamp_3 matlab.ui.control.Lamp

end

properties (Access = private)

imds0;

table;

imds1;

NFrame;

frames;

Name;

net;

predictY;

Info;

% Description

end

methods (Access = private)

end

% Callbacks that handle component events

methods (Access = private)

% Code that executes after component creation

function startupFcn(app, NET, example)

app.net=load("net.mat");

example=imageDatastore('example.png');

img = readimage(example,1);

imshow(img,'Parent',app.PLOT);

app.table = readtable('Example.xlsx');

app.UITable.Data=app.table;

end

% Value changed function: Load

function LoadValueChanged(app, event)

DataLocation=uigetdir('Location of the data');

app.imds0 = imageDatastore(fullfile(DataLocation),'FileExtensions','.png');

app.imds1 = augmentedImageDatastore([224 224],app.imds0 );

% N° of frames

S=size(app.imds0.Files);

app.frames=S(1,1);

app.NFrame = 1:1:S(1,1);app.NFrame=app.NFrame';

app.NFrame;

app.table.FrameN_(1:S(1,1))=app.NFrame;

% Names of those frames

app.Name=app.imds0.Files;

app.table.ImageName(1:S(1,1))=app.Name;

% Load all the data

app.UITable.Data=app.table;

% Switch on lamp

app.DoneLamp.Color='g';

end

% Value changing function: FrameNSpinner

function FrameNSpinnerValueChanging(app, event)

end

% Value changed function: FrameNSpinner

function FrameNSpinnerValueChanged(app, event)

FrameN = app.FrameNSpinner.Value;

images1=readimage(app.imds0,FrameN);

imshow(images1,'Parent',app.PLOT)

end

% Value changed function: Predict

function PredictValueChanged(app, event)

app.net

app.imds1

group=18;

[YPred]= classify(app.net,app.imds1);

s=string(YPred);

app.predictY=double(s);

app.table.Distance(1:app.frames)=app.predictY;

app.DoneLamp_2.Color='g';

end

end

% Component initialization

methods (Access = private)

% Create UIFigure and components

function createComponents(app)

% Create UIFigure and hide until all components are created

app.UIFigure = uifigure('Visible', 'off');

app.UIFigure.Position = [100 100 1170 742];

app.UIFigure.Name = 'UI Figure';

% Create DistanceEvolution

app.DistanceEvolution = uiaxes(app.UIFigure);

title(app.DistanceEvolution, 'Distance evolution ')

xlabel(app.DistanceEvolution, 'Frame N°')

ylabel(app.DistanceEvolution, 'Distance(pixels)')

app.DistanceEvolution.Position = [561 40 570 321];

% Create UITable

app.UITable = uitable(app.UIFigure);

app.UITable.ColumnName = {'Frame N°'; 'ImageName'; 'Distance'};

app.UITable.ColumnWidth = {'auto'};

app.UITable.RowName = {};

app.UITable.Position = [17 73 528 274];

% Create Load

app.Load = uibutton(app.UIFigure, 'state');

app.Load.ValueChangedFcn = createCallbackFcn(app, @LoadValueChanged, true);

app.Load.Text = 'LOAD';

app.Load.FontWeight = 'bold';

app.Load.Position = [91 611 100 22];

% Create Predict

app.Predict = uibutton(app.UIFigure, 'state');

app.Predict.ValueChangedFcn = createCallbackFcn(app, @PredictValueChanged, true);

app.Predict.Text = {'PREDICT'; ''};

app.Predict.FontWeight = 'bold';

app.Predict.Position = [91 581 100 22];

% Create FrameNSpinnerLabel

app.FrameNSpinnerLabel = uilabel(app.UIFigure);

app.FrameNSpinnerLabel.HorizontalAlignment = 'center';

app.FrameNSpinnerLabel.Position = [699 404 57 22];

app.FrameNSpinnerLabel.Text = 'Frame N°';

% Create FrameNSpinner

app.FrameNSpinner = uispinner(app.UIFigure);

app.FrameNSpinner.ValueChangingFcn = createCallbackFcn(app, @FrameNSpinnerValueChanging, true);

app.FrameNSpinner.ValueChangedFcn = createCallbackFcn(app, @FrameNSpinnerValueChanged, true);

app.FrameNSpinner.HorizontalAlignment = 'center';

app.FrameNSpinner.Position = [755 404 45 22];

app.FrameNSpinner.Value = 1;

% Create Save

app.Save = uibutton(app.UIFigure, 'push');

app.Save.FontWeight = 'bold';

app.Save.Position = [91 551 100 22];

app.Save.Text = {'SAVE'; ''};

% Create DistanceextrantionappLabel

app.DistanceextrantionappLabel = uilabel(app.UIFigure);

app.DistanceextrantionappLabel.FontSize = 40;

app.DistanceextrantionappLabel.Position = [111 685 434 48];

app.DistanceextrantionappLabel.Text = 'Distance extrantion app';

% Create Image

app.Image = uiimage(app.UIFigure);

app.Image.Position = [17 653 86 80];

app.Image.ImageSource = '0.jpg';

% Create Image_2

app.Image_2 = uiimage(app.UIFigure);

app.Image_2.Position = [544 653 86 80];

app.Image_2.ImageSource = 'ijm_logo.jpg';

% Create PLOT

app.PLOT = uiaxes(app.UIFigure);

title(app.PLOT, 'Preview')

xlabel(app.PLOT, '')

ylabel(app.PLOT, '')

app.PLOT.FontSize = 14;

app.PLOT.Position = [799 360 332 344];

% Create DoneLampLabel

app.DoneLampLabel = uilabel(app.UIFigure);

app.DoneLampLabel.HorizontalAlignment = 'right';

app.DoneLampLabel.Position = [209 611 34 22];

app.DoneLampLabel.Text = 'Done';

% Create DoneLamp

app.DoneLamp = uilamp(app.UIFigure);

app.DoneLamp.Position = [258 611 20 20];

app.DoneLamp.Color = [1 0 0];

% Create DoneLamp_2Label

app.DoneLamp_2Label = uilabel(app.UIFigure);

app.DoneLamp_2Label.HorizontalAlignment = 'right';

app.DoneLamp_2Label.Position = [209 581 34 22];

app.DoneLamp_2Label.Text = 'Done';

% Create DoneLamp_2

app.DoneLamp_2 = uilamp(app.UIFigure);

app.DoneLamp_2.Position = [258 581 20 20];

app.DoneLamp_2.Color = [1 0 0];

% Create DoneLamp_3Label

app.DoneLamp_3Label = uilabel(app.UIFigure);

app.DoneLamp_3Label.HorizontalAlignment = 'right';

app.DoneLamp_3Label.Position = [209 551 34 22];

app.DoneLamp_3Label.Text = 'Done';

% Create DoneLamp_3

app.DoneLamp_3 = uilamp(app.UIFigure);

app.DoneLamp_3.Position = [258 551 20 20];

app.DoneLamp_3.Color = [1 0 0];

% Show the figure after all components are created

app.UIFigure.Visible = 'on';

end

end

% App creation and deletion

methods (Access = public)

% Construct app

function app = ICapp(varargin)

% Create UIFigure and components

createComponents(app)

% Register the app with App Designer

registerApp(app, app.UIFigure)

% Execute the startup function

runStartupFcn(app, @(app)startupFcn(app, varargin{:}))

if nargout == 0

clear app

end

end

% Code that executes before app deletion

function delete(app)

% Delete UIFigure when app is deleted

delete(app.UIFigure)

end

end

end

Answer 1

[Kojiro Saito]

Your classify function is treated as that of Statistics and Machine Learning Toolbox in the compiled application. In order to force compiled application to use CNN classify, there are two ways.

(1) Use SeriesNetwork.loadobj

function startupFcn(app, NET, example)
  app.net = load("net.mat");
  app.net = SeriesNetwork.loadobj(app.net.net);
  % ... %
end 

(2) User function progma (%#) in the first line of startupFcn

function startupFcn(app, NET, example)
  %#function SeriesNetwork
  app.net = load("net.mat");
  app.net = app.net.net;
  % ... %
end

Hope this helps..

[Pablo Salaverria]

It worked!!!! Thank you very much!!!

[Shawn Dante]

i am having an error at app.net.net

it says "Reference to non existent field 'net'.

[Kojiro Saito]

@Shawn Dante

Does your mat file contain variable net? app.net.net should be app.net.a if your mat file contains variable a.

[Shawn Dante]

@Kojiro Saito Thank You Sir it worked i am also having an problem at label=classify(app.net,app.myimage); error:Undefined function classify for input argument of type unit 8

//app.myimage is a single test image file

[Kojiro Saito]

@Shawn Dante

What is your app.net model? SeriesNetwork, DAGNetwork or other?

Does classify run fine by command line?

[Shawn Dante]

@Kojiro Saito

its a CNN model, classify does not run in my command line

[Kojiro Saito]

@Shawn Dante

Have you installed Deep Learning Toolbox on your MATLAB?

Does ver command return Deep Learning Toolbox?

ver

[Shawn Dante]

yes it contains deep learning toolbox

[Kojiro Saito]

@Shawn Dante

OK, could you try the following? Please replace yourimage.jpg to your actual image file name and extension.

app.myimage = imageDatastore('yourimage.jpg');   
label=classify(app.net,app.myimage);   

[Shawn Dante]

@Kojiro Saito

again it gives the same error error:Undefined function classify for input argument of type matlab.io.datastore.ImageDatastore'.

[Kojiro Saito]

@Shawn Dante

Maybe you have another classify.

Does

which classify -all

returns properly toolbox\nnet\cnn\@DAGNetwork\classify.m?

[Shawn Dante]

@Kojiro Saito

yes it returns

[Kojiro Saito]

@Shawn Dante

So, the cause might be your mat file.

After loading the mat file,

whos

returns DAGNetwork or SeriesNetwork in the Class column?

[Shawn Dante]

@Kojiro Saito

its Series network

[Kojiro Saito]

@Shawn Dante

OK, thank you. So it would work...

Have you tried other models, for example, AlexNet?

app.net = alexnet;
label=classify(app.net,app.myimage);

[Shawn Dante]

yes it works for alexnet

[Kojiro Saito]

@Shawn Dante

OK, so alexnet works. Could you upload or share your model (.mat)?

[Kojiro Saito]

@Shawn Dante

Thanks. I finally understand that this is due to the mat file.

For creating mat file which contains a deep learning model, I would do the following.

save('csnet.mat', 'csnet')

If so, csnet.mat only contains variable csnet.

But your mat file contains other variables like I and img.

To work with existing csnet.mat, you need the following.

(1) In your startupFcn function, app.net should be csnet variable.

function startupFcn(app)
    app.net=load("csnet.mat");
    %app.net=app.net.I;
    app.net=app.net.csnet;
    %...%
end

(2) In your ResultButtonPushed function, it's not possible to merge label and app.myimage, so try insertText instead.

function ResultButtonPushed(app, event)
    label=classify(app.net,app.myimage);
    %imshow([label,app.myimage],'Parent',app.UIAxes_2)
    myimage2 = insertText(app.myimage, [0 5], string(label),'FontSize', 24);
    imshow(myimage2, 'Parent',app.UIAxes_2)
end

[fariha Shahid]

I am trying to load my net.mat file, which I have created using nnstart command, but it is giving me error

Error using load
Unable to read file 'net.mat'. No such file or directory.

[Kristie Nguyen]

@Kojiro Saito

Can you explain what to do when the network I am trying to load into matlab is a DAGnetwork?

methods (Access = private)

        % Code that executes after component creation
        function startupFcn(app)
            
            app = load('net.mat');
            app.net = app.net.net;
        end

        % Image clicked function: Image
        function ImageClicked(app, event)
            file = uigetfile("*.jpg");
            test = imread(file);
            img = imresize(test,[224 224]);
            app.Image.ImageSource = img;
            pred = classify(app.net,img);
            app.Label.Labelsource = pred;
        end
    end

[Kojiro Saito]

@Kristie Nguyen

function startupFcn(app)
    %#exclude net.mat
    %#function DAGNetwork
    app = load('net.mat');
    ...
end

For detail, please refer to these documents (function pragma and exclude pragma).

[仲富 刘]

i meet these problems too using matlab2020 and i just want to use it to classify my picture,what should i do

[仲富 刘]

can you give me a help ,i save my picture as a global data,code like this :

as start: app=load('net.mat');

app.net = Net;%Net is the name of the net data.

the classify function is:

global im; reim=imresize(im,[227 227]);

[result,score] = classify(app.net,reim);

there are a lot of problems about "net",what should i do:

[仲富 刘]

it says :The property 'net' of class 'app' is not recognized

[仲富 刘]

It solved by adding a property! Thank you!

[Fatin Nasuha Bt Asrol]

Hi..can I get some help to look at my coding..I still get error for the classification result button..Hopefully someone can help me.

This is my coding:

• 

[Kojiro Saito]

@Fatin Nasuha Bt Asrol

Did you add function pragma (starts with %#) in startupFcn? This is not just a comment but a pragma.

function startupFcn(app)
    %#function SeriesNetwork
    app.net = load("net.mat");
    app.net = app.net.net;
end

If this does not work, try DAGNetwork function pragma.

function startupFcn(app)
    %#function DAGNetwork
    app.net = load("net.mat");
    app.net = app.net.net;
end

[Fatin Nasuha Bt Asrol]

@Kojiro Saito Thank you so much for your reply. I already try DAGNetwork function pragma.However, I still get the same error.

Do you know what is the possible cause for this? will much appreaciated if you can assist me on this.Thank you.

[Kojiro Saito]

@Fatin Nasuha Bt Asrol

It's better to check actual class.

Does the following return DAGNetwork in Class?

whos app.net

[Fatin Nasuha Bt Asrol]

@Kojiro Saito No, it did not return DAGNetwork in Class.Did you know how to make it happen?

[Kojiro Saito]

@Fatin Nasuha Bt Asrol

What variables are contained in your mat file?

The following example shows DAGNetwork.

net = googlenet;
save net

% Suppose reading mat file in AppDesigner
app.net = load('net.mat');

[Fatin Nasuha Bt Asrol]

@Kojiro Saito Thank you for your reply. However I still get the same error. Actually, I used result trainedNetwork _1 in deep network designer and save it as net.mat file.

As the image below:

This is in app designer:

And to classify the images I reused it in app designer as the image above. Is it my method correct to classify image in app designer?

[Kojiro Saito]

@Fatin Nasuha Bt Asrol

You need to extract trainedNetwork_1 variable from net.mat file.

function startupFcn(app)
    %#function DAGNetwork
    app.net = load('net.mat');
    app.net = app.net.trainedNetwork_1;
end

function ClassificationresultButtonPushed(app, event)
    [YPred, probs] = classify(app.net, I);
end

[Fatin Nasuha Bt Asrol]

@Kojiro Saito Thank you for your reply. I already tried your suggestion extract trainedNetwork_1 variable from net.mat file but I still get the same error [ Undefined function 'classify' for input arguments of type 'uint8']. Is it another way to solve it?

[Kojiro Saito]

@Fatin Nasuha Bt Asrol

OK, apart from App Designer and MATLAB Compiler, let's execute classify in MATLAB command window.

What did you get by the following command?

load('net.mat')
img = imread('somefile.jpg');
imgRes = imresize(img, [244 244]);
[YPred, probs] = classify(trainedNetwork_1, imgRes);

[Fatin Nasuha Bt Asrol]

@Kojiro Saito Thank you for reply, I got this error:

[Kojiro Saito]

@Fatin Nasuha Bt Asrol

I copied your code, but the error says imresize to 244x244 is wrong.

Fix it by the following.

imgRes = imresize(img, [224 224]);

[Fatin Nasuha Bt Asrol]

@Kojiro Saito Thank you for reply. I already tried change it according your coding imgRes = imresize (img, [244 244]); but I get the another error like this:

[Kojiro Saito]

@Fatin Nasuha Bt Asrol

I'm a bit confused.

Is Deep Learning Toolbox installed and you have a valid license?

Is your trained model (trainedNetwork_1) allows imageInputLayer?

Your question is not related to App Designer any more, so if the above does not work, I think it's better if you would post a question in MATLAB Answers as a new question.