Use Zed captured data for MediaPipe and SMPL-X human model

Hello, I would like to ask how could I use Zed captured data to apply MediaPipe for extracting human face keypoints and then apply the extracted keypoints (from MediaPipe and Zed SDK) to SMPL-X human model?

Any solution? I found this paper: Real-time RGBD-based Extended Body Pose Estimation, but it works on Azure Kinect SDK