Coordinates of output (x,y) tracking from cropped videos

Hi everyone,
I’m using deeplabcut 2.0.4.1 in ubuntu 16.04:
My question is: After tracking (x,y) positions from cropped videos, the output (x,y) positions will be in the coordinates of my original video or in the coordinates of the cropped region?
thanks…

It will be in the coordinates of the cropped video, but you can compute it from the cropping variables in the pickle file (see contents here):
https://github.com/AlexEMG/DeepLabCut/blob/master/deeplabcut/pose_estimation_tensorflow/predict_videos.py#L292