VideoPose3d_with_Detectron2 icon indicating copy to clipboard operation
VideoPose3d_with_Detectron2 copied to clipboard

Published 20 hours ago • verified publisher icon darkAlert

Function encode_for_videpose3d

Open laumc98 opened this issue 4 years ago • 2 comments

laumc98 avatar Mar 27 '20 21:03 laumc98

Hi,I'm currently working with detectron2 and I'm trying to get the information from the boxes and the keypoints detected. In your function encode_for_videpose3d the variables that you enter (boxes,keypoints)) are from the outputs of the predictor ?

laumc98 avatar Mar 27 '20 21:03 laumc98

@laumc98 Hi, yes, they are predicted by the detectron2 (COCO Person Keypoint R-CNN with 'X101-FPN' backbone)

darkAlert avatar Mar 28 '20 10:03 darkAlert