Using posenet for estimating, 2d detections and input it to 3d pose lifter

DenisTome / Lifting-from-the-Deep-release

Implementation of "Lifting from the Deep: Convolutional 3D Pose Estimation from a Single Image"

https://denistome.github.io/papers/lifting-from-the-deep

GNU General Public License v3.0

450 stars 132 forks source link

Using posenet for estimating, 2d detections and input it to 3d pose lifter #39

Closed timtensor closed 5 years ago

timtensor commented 5 years ago

Hi , I am currently looking into python implementationo posenet to do 2D pose estimation.
posenet-python. In it , the following decode single person. The following return the 17 keypoints instance_keypoint_coords and pose confidence. instance_keypoint_scores .

Where can we input to get the corresponding 3d estimation ?

Can it also be extended to videos ?

DenisTome commented 5 years ago

I'm not sure I get exactly what you are trying to do. If by that you mean that you want to use their 2D estimations and this repository to lift them in 3D, then at line 144 of packages/lifting/_pose_estimator.py use can see where the pose + visibility is given to the 3D lifting module. Just make sure that the format of the pose is the same as the one expected by our approach.