Problem with detectSURFFeatures in bagOfFeatures

Hello, it's my first post, so sorry if I set something not correctly.

I'm trying to write my own version of Image Recognition Using Machine Learning. I'm using code from this video: https://www.mathworks.com/videos/image-recognition-using-machine-learning-122900.html. What I change is image files (I use only two categories - drones and birds and I have only 200 examples from each group).

After running this section:

    %%Test out accuracy on test set!
testSceneData = double(encode(bag, test_set));
testSceneData = array2table(testSceneData,'VariableNames',trainedClassifier.RequiredVariables);
actualSceneType = test_set.Labels;
predictedOutcome = trainedClassifier.predictFcn(testSceneData);
correctPredictions = (predictedOutcome == actualSceneType);
validationAccuracy = sum(correctPredictions)/length(predictedOutcome) %#ok

I have error like this:

   Error using detectSURFFeatures
Expected input number 1, I, to be one of these types:
logical, uint8, int16, uint16, single, double
Instead its type was matlab.io.datastore.ImageDatastore.
Error in detectSURFFeatures>checkImage (line 124)
validateattributes(I,{'logical', 'uint8', 'int16', 'uint16', ...
Error in detectSURFFeatures (line 81)
checkImage(I);
Error in bagOfFeatures/determineExtractionPoints (line 591)
                points = detectSURFFeatures(grayImage, 'MetricThreshold', 800);
Error in bagOfFeatures/extractDescriptorsFromImage (line 645)
            points = this.determineExtractionPoints(grayImage);
Error in bagOfFeatures>@(img)extractDescriptorsFromImage(this,img)
Error in bagOfFeatures/encodeSingleImage (line 487)
                descriptors = this.Extractor(img);
Error in bagOfFeatures/encode (line 355)
                [featureVector, varargout{1:numVarargout}] = this.encodeSingleImage(in, params);

I create variable bag in this way:

    bag = bagOfFeatures(imageSet(training_set.Files),'VocabularySize',250,'PointSelection','Detector');

Do you have any idea what can I change to fix this problem?

Answer 1

[Jakub Ajchel]

Correct solution:

testSceneData = double(encode(bag, imageSet(test_set.Files)));

test_set was an imageDatastore. This is specific type of data.

Function encode requires array of images. That's the reason why we should convert test_set into imageSet, which is an array of images