Cant use more than one class, strcmp issue?

I am trying to implement more labels into my YoloV4 CNN, but I am getting this error message.

I am not really sure what is going on, I have tried to go into the problem areas and find something out, but no luck. not really sure what the gTruthClassNames is, but I think thats my problem. I would attach my code but it is to large (any way around that)

Could this also be caused by not using one of the class names not being used, is there a way around this?

Here is my class names

className = {'Front-Windscreen-Damage'...
,'Headlight-Damage'...
,'Major-Rear-Bumper-Dent'...
,'Rear-windscreen-Damage'...
,'RunningBoard-Dent'...
,'Sidemirror-Damage'...
,'Signlight-Damage'...
,'Taillight-Damage'...
,'bonnet-dent'...
,'doorouter-dent'...
,'fender-dent'...
,'front-bumper-dent'...
,'medium-Bodypanel-Dent'...
,'pillar-dent'...
,'quaterpanel-dent'...
,'rear-bumper-dent'...
,'roof-dent'} ;

These are the imputs, Ill share a link to my workspace below, it goes to my google drive as it is over 5mb. (Not sure if sharing like this is allowed, remove is not)...

[detector,info] = trainYOLOv4ObjectDetector(augmentedTrainingData,detector,options);

https://drive.google.com/file/d/16ZTtyE5Yb6U5iMzkaG41a4gjNee6_aDK/view?usp=sharing

Thanks everyone!

Answer 1

[Voss]

This error happens because the inputs to strcmp are incompatible sizes.

Example 1: both inputs the same size (works):

a = {'baba', 'booey'};
b = {'barbara', 'booey'};
strcmp(a,b) % works

ans = 1×2 logical array
   0   1

Example 2: one input a scalar, one non-scalar (works):

a = {'baba'};
b = {'barbara', 'booey'};
strcmp(a,b) % works

ans = 1×2 logical array
   0   0

Example 3: non-scalar inputs of different sizes (error):

a = {'baba', 'booey', 'mountain'};
b = {'barbara', 'booey'};
strcmp(a,b) % error

Error using strcmp
Inputs must be the same size or either one can be a scalar.

This means that, in your case, gTruthClassNames and detectorClassNames are non-scalars with different sizes. I'm not able to say why that is or what you should do about it though.

[Conner Carriere]

Ok, so since mine will not be scalars, I have to pull those two values out and find a way to make them the same size.