Expected LABEL to be nonempty error when running YOLOv2 on my own dataset

hello,

when i try to run YOLOv2 as in https://www.mathworks.com/help/deeplearning/examples/object-detection-using-yolo-v2.html

using my own dataset i got this kind of error and i don't know what;s wrong with my dataset? could anyone please explian what does this erorr mean and how can in solve it ?

Error using insertObjectAnnotation

Expected LABEL to be nonempty.

Error in insertObjectAnnotation>checkLabel (line 290)

validateattributes(label, {'numeric'}, ...

Error in insertObjectAnnotation>validateAndParseInputs (line 180)

checkLabel(label);

Error in insertObjectAnnotation (line 123)

validateAndParseInputs(I, shape, position, label, varargin{:});

Error in YOLOv2 (line 153)

I = insertObjectAnnotation(I,'rectangle',bboxes,scores);

[Nicolai Valenti]

Hi, I got the same error but with the standard dataset, have you managed to resolve the issue?

[raghad Abdulaziz]

check this https://www.mathworks.com/matlabcentral/answers/389238-what-is-wrong-in-this-code

when I increased my dataset it worked

and in other run I also increased the number of epochs

good luck!

[Yonathan Zarkovian]

Hi, did you figure out the problem?

I'm getting the same error when following the yolo v3 guide. It only happens with certain images.

Does the error mean that it couldn't detect an object, or is it something else?

Thanks.

[Dave Eccles]

Was this issue resolved? I also seem to be receiving the same error code with a new image dataset.

[Dimple Sethi]

I am facing the same error with certain videos. But for rest of the videos, its working fine. Videos are almost same. Please help to resolve the issue,

[ybec]

I am getting the same error, but it appears when the image "highway.png" is changed for another one.

I = imread('highway.png');
I = imresize(I,inputSize(1:2));
[bboxes,scores] = detect(detector,I);

[ybec]

Yonathan Zarkovian, I think the error shows up when an image is not what the algorithm is waiting for. I uploaded different kind of images and only similar images to the dataset worked.