3D body reconstruction from a single camera

result
This video shows how I reconstruct the 3D skeleton pose, the body shape and the texture of myself based on the video from a single camera. We are still working on it to better simulate the face, hand and foot.

The following picture shows the settings for our demo. I’m the person on the left in the Third Person View, and I’m wearing the collaboration headset recording my friend playing Tai Chi. My code will try to update and refine the texture and body shape for the target person. The reconstruction result shown on the left is based only on my front view camera, two downward looking cameras and a few inertial sensors. demo more details are comming soon…

I’m seeking a 2020 summer internship position in AR/VR. Please let me know if you are interested in my research. My email is me[at]qzane.com

Written on October 28, 2019