Error while training SSD object detector

12 views (last 30 days)
i am using MATLAB 2020a to train SSD object detetctor for my database.I receive error which i am unable to troubleshoot. please help me out.
here is the code:
vehicleDataset1=vehicleDataset1.wbc;
rng(0);
shuffledIndices1 = randperm(height(vehicleDataset1));
idx1 = floor(0.6 * length(shuffledIndices1) );
trainingData1 = vehicleDataset1(shuffledIndices1(1:idx1),:);
testData1 = vehicleDataset1(shuffledIndices1(idx1+1:end),:);
imdsTrain1 = imageDatastore(trainingData1{:,'filename'});
bldsTrain1 = boxLabelDatastore(trainingData1(:,'wbc'));
imdsTest1 = imageDatastore(testData1{:,'filename'});
bldsTest1 = boxLabelDatastore(testData1(:,'wbc'));
trainingData1 = combine(imdsTrain1,bldsTrain1);
testData1 = combine(imdsTest1, bldsTest1);
data1 = read(trainingData1);
I1 = data1{1};
bbox1 = data1{2};
annotatedImage1 = insertShape(I1,'Rectangle',bbox1);
annotatedImage1 = imresize(annotatedImage1,2);
figure
imshow(annotatedImage1)
inputSize1 = [300 300 3];
numClasses1 = width(vehicleDataset1)-1;
lgraph1 = ssdLayers(inputSize1, numClasses1, 'resnet50');
augmentedTrainingData1 = transform(trainingData1,@augmentData);
augmentedData1 = cell(4,1);
for k = 1:4
data1 = read(augmentedTrainingData1);
augmentedData1{k} = insertShape(data1{1},'Rectangle',data1{2});
reset(augmentedTrainingData1);
end
figure
montage(augmentedData1,'BorderSize',10)
preprocessedTrainingData1 = transform(augmentedTrainingData1,@(data)preprocessData(data,inputSize1));
data1 = read(preprocessedTrainingData1);
I1 = data1{1};
bbox1 = data1{2};
annotatedImage1 = insertShape(I1,'Rectangle',bbox1);
annotatedImage1 = imresize(annotatedImage1,2);
figure
imshow(annotatedImage1)
options = trainingOptions('sgdm',...
'InitialLearnRate',5e-5,...
'MiniBatchSize',16,...
'Verbose',true,...
'MaxEpochs',50,...
'Shuffle','every-epoch',...
'VerboseFrequency',10,...
'CheckpointPath',tempdir);
[detector1,info1] = trainSSDObjectDetector(preprocessedTrainingData1,lgraph1,options);
data1 = read(testData1);
I1 = data1{1,1};
I1 = imresize(I1,inputSize1(1:2));
[bboxes1,scores1] = detect(detector1,I1, 'Threshold', 0.4);
I1 = insertObjectAnnotation(I1,'rectangle',bboxes1,scores1);
figure
imshow(I1)
here is error:
Invalid transform function defined on datastore.
The cause of the error was:
Error using bboxwarp>iParseInputs (line 324)
The value of 'bboxA' is invalid. Expected input number 1, bboxA, to be integer-valued.
Error in bboxwarp (line 81)
params = iParseInputs(bboxA,tform,ref,varargin{:});
Error in augmentData (line 23)
[B{2},indices] = bboxwarp(A{2},tform,rout,'OverlapThreshold',0.25);
Error in matlab.io.datastore.TransformedDatastore/applyTransforms (line 489)
data = ds.Transforms{ii}(data);
Error in matlab.io.datastore.TransformedDatastore/read (line 162)
[data, info] = ds.applyTransforms(data, info);
Error in nnet.internal.cnn.DataLoader/manageReadQueue (line 161)
data = read(self.Datastore);
Error in nnet.internal.cnn.DataLoader/readAhead (line 192)
manageReadQueue(self);
Error in nnet.internal.cnn.DataLoader (line 80)
readAhead(self);
Error in nnet.internal.cnn.GeneralDatastoreDispatcher (line 272)
this.DataLoader = nnet.internal.cnn.DataLoader(ds,...
Error in nnet.internal.cnn.DataDispatcherFactory.createDataDispatcherMIMO (line 170)
nnet.internal.cnn.GeneralDatastoreDispatcher( ...
Error in vision.internal.cnn.trainNetwork>iCreateTrainingDataDispatcher (line 190)
dispatcher = nnet.internal.cnn.DataDispatcherFactory.createDataDispatcherMIMO( ...
Error in vision.internal.cnn.trainNetwork (line 40)
trainingDispatcher = iCreateTrainingDataDispatcher(ds, mapping, trainedNet,...
Error in trainSSDObjectDetector (line 233)
[network, info] = vision.internal.cnn.trainNetwork(...
Error in wbc_detetctor_traininga (line 61)
[detector1,info1] = trainSSDObjectDetector(preprocessedTrainingData1,lgraph1,options);

Accepted Answer

Divya Gaddipati
Divya Gaddipati on 16 Jun 2020
Make sure your groundtruths are valid and non-empty i.e., the values of the bounding boxes are finite, positive, non-fractional, non-NaN and should be within the image boundary with a positive height and width. You should either discard or fix the samples with invalid bounding boxes.
  2 Comments
Huma Hafeez
Huma Hafeez on 24 May 2021
Thanks for your reply. You are right, there was a problem with ground truth

Sign in to comment.

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!