How to fix error from evaluatesemanticsegmentation function?

Question

yasmin ismail 2023-10-4

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2029279-how-to-fix-error-from-evaluatesemanticsegmentation-function

评论： Walter Roberson 2023-12-14

在 MATLAB Online 中打开

I used the code for evaluation semantic segmentation in

Evaluate semantic segmentation data set against ground truth - MATLAB evaluateSemanticSegmentation (mathworks.com)

as shown below:

clc
clear all
dataSetDir = fullfile(toolboxdir('vision'),'visiondata','triangleImages');
testImagesDir = fullfile(dataSetDir,'testImages');
testLabelsDir = fullfile(dataSetDir,'testLabels');
imds = imageDatastore(testImagesDir);
classNames = ["crack","background"];
labelIDs   = [255 0];
pxdsTruth = pixelLabelDatastore(testLabelsDir,classNames,labelIDs);
data = load('triangleSegmentationNetwork');
net = data.net
net = 
  SeriesNetwork with properties:

         Layers: [10×1 nnet.cnn.layer.Layer]
     InputNames: {'imageinput'}
    OutputNames: {'classoutput'}
pxdsResults = semanticseg(imds,net,"WriteLocation",tempdir);
Running semantic segmentation network
-------------------------------------
* Processed 100 images.
metrics = evaluateSemanticSegmentation(pxdsResults,pxdsTruth)
Evaluating semantic segmentation results
----------------------------------------
* Selected metrics: global accuracy, class accuracy, IoU, weighted IoU, BF score.
* Processed 0 images.
Error using semanticSegmentationMetrics>iAssertCategoricalsHaveSameCategories
The categorical data returned by dsResults and dsTruth must have the same categories.

Error in semanticSegmentationMetrics/processImages (line 589)
                iAssertCategoricalsHaveSameCategories(act,exp,obj.ClassNames);

Error in semanticSegmentationMetrics/processImagesSerial (line 663)
                    m = processImages(obj,data);

Error in semanticSegmentationMetrics/computeMetrics (line 389)
                    metrics = processImagesSerial(obj);

Error in semanticSegmentationMetrics (line 195)
            [obj, blockMetrics] = obj.computeMetrics();

Error in semanticSegmentationMetrics.compute (line 174)
            [obj, blockMetrics] = semanticSegmentationMetrics(varargin{:});

Error in evaluateSemanticSegmentation (line 205)
[metrics, blockMetrics, canComputeBlockMetrics] = semanticSegmentationMetrics.compute(varargin{:});

I got the following error:

Error using semanticSegmentationMetrics>iAssertCategoricalsHaveSameCategories (line 1226)

The categorical data returned by dsResults and dsTruth must have the same categories.

Error in semanticSegmentationMetrics/processImages (line 589)

iAssertCategoricalsHaveSameCategories(act,exp,obj.ClassNames);

Error in semanticSegmentationMetrics/processImagesSerial (line 663)

m = processImages(obj,data);

Error in semanticSegmentationMetrics/computeMetrics (line 389)

metrics = processImagesSerial(obj);

Error in semanticSegmentationMetrics (line 195)

[obj, blockMetrics] = obj.computeMetrics();

Error in semanticSegmentationMetrics.compute (line 174)

[obj, blockMetrics] = semanticSegmentationMetrics(varargin{:});

Error in evaluateSemanticSegmentation (line 205)

[metrics, blockMetrics, canComputeBlockMetrics] = semanticSegmentationMetrics.compute(varargin{:});

Error in Gtruth2 (line 13)

metrics = evaluateSemanticSegmentation(pxdsResults,pxdsTruth)

I noted that I have different categories between Result and Truth, but I don’t know how to fix it. I named the attached file same as mentioned in the code on Evaluate semantic segmentation data set against ground truth - MATLAB evaluateSemanticSegmentation (mathworks.com) (triangleImages)

the attached (R7001-60.png) is from testImages folder , the attached (gT7001_60.png) is from testLables folder , both folders in (triangleImages) folder.

How to fix it?