vita-epfl / openpifpaf

Official implementation of "OpenPifPaf: Composite Fields for Semantic Keypoint Detection and Spatio-Temporal Association" in PyTorch.
https://arxiv.org/abs/2103.02440
Other
80 stars 22 forks source link

Question : Can u please tell, what is the format of json in the keypoint and the bounding box is being stored ? #23

Open Chaitanya887 opened 2 weeks ago

Chaitanya887 commented 2 weeks ago

image For the bounding box, is it the yolo format which is

center x-coordinate ,center y-coordinate ,height ,width

or is it coordinates of upper left (x1, y1)and lower right (x2, y2) part of the bounding box

x1, y1, x2, y2

Also can u tell the same about the keypoint

yasaminborhani commented 3 days ago

Hello,

The bounding box format is x_min, y_min, w, h.

For keypoint format please check this website: https://vita-epfl.github.io/openpifpaf/plugins_custom.html There are some explanations there to help you implement your data loader. If you want to use the datasets which their data loaders are already available, you can easily use the data modules and the appropriate format will be produced as the output.