Using my own dataset how to define the keypoints in json file?

[esteimle]

Hi, I got this running and training with the ai_challenger data and now I'd like to try some of my own data. Bu I don't quite understand the whole format for the ai_challenger_train.json file. For example in that file there are annotations, under annotations I find elements for each image. In those elements there is then a num_keypoints which seems to vary and I don't know why it isn't always 14 (some not visible maybe?). Then there is a keypoints list which is always 42 items long which looks to be 42/3 = 14 so I assume there are three items for each keypoint. For example:

keypoints->[0] |->432 |->234 |->2

I assume it goes: |->x |->y |->??

I don't know what the last one, the 2 means. Is it a category or a depth of some kind or an indicator of what is visible or not? I'm really confused here. I think if I could figure out the formatting I could start training my own data and this is the last thing holding me up. Thank you for any help you can give.

[esteimle]

ah nm I think I get it after plotting points on pictures. 1 is occluded, 2 is visible and num_keypoints is the number of visible points. That's my guess anyway.