How should my region-based training json format be?

[Cestovatels]

Hello,

I want to provide training on region based segmentation for tooth numbering segmentation.
There are 32 classes for tooth number and one class for pulp.
The pulp class will be in almost all tooth segmentations, maybe some teeth may not contain the pulp class.
I set dataset.json as follows.
Is this correct or not?
All teeth have different positions, they don't contain each other, only all of them contain the pulp class.
Can I use region based traning in this case?

...
"labels": {
    "background": 0,
    "teeth_1": [1,33],
    "teeth_2": [2, 33],
    "teeth_3": [3,33],
      ...
      ...
      ...
     "teeth_32": [32,33],
     "pulp": [33],
},
"regions_class_order": [1, 2, 3, 5, 6, ... 33],
...

[Cestovatels]

Hello,

I want to provide training on region based segmentation for tooth numbering segmentation. There are 32 classes for tooth number and one class for pulp. The pulp class will be in almost all tooth segmentations, maybe some teeth may not contain the pulp class. I set dataset.json as follows. Is this correct or not? All teeth have different positions, they don't contain each other, only all of them contain the pulp class. Can I use region based traning in this case?

...
"labels": {
    "background": 0,
    "teeth_1": [1,33],
    "teeth_2": [2, 33],
    "teeth_3": [3,33],
      ...
      ...
      ...
     "teeth_32": [32,33],
     "pulp": [33],
},
"regions_class_order": [1, 2, 3, 5, 6, ... 33],
...

Is the json I wrote about the subject correct? I would be very grateful if you answer.
@mrokuss @FabianIsensee