You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
<papertitle>Realtime 3D Deep Motion Capture</papertitle>
246
247
</a>
247
248
<br>
248
249
Oct' 20 – Dec' 20
249
250
<br>
250
251
<br>
251
-
To Complete
252
+
In collaboration with Pierre Tessier (MS, Columbia University), the objective of our project was to implement a intelligent 2D to 3D Motion Capture mechanism that uses only the video stream of a webcam as input. We were able to animate relatively accurately this Mk-44 Iron Man 3D model.
252
253
<br>
253
254
<br>
254
-
<p></p>
255
-
255
+
The project is based on the model DOPE presented in <aclass="tog" href="https://arxiv.org/abs/2008.09457" target="_blank">this paper</a> for the automatic 3D rig generation from video input coupled with a <aclass="tog" href="https://www.researchgate.net/publication/230600780_Closed-Form_Solution_of_Absolute_Orientation_Using_Unit_Quaternions" target="_blank">quaternion</a>-based 3D rotation inference pipeline for 3D model animation.
Looking back at this project, I saw that <aclass="tog" href="https://arxiv.org/abs/1504.03504" target="_blank">another paper</a> with much more impressive results came out a few years later using <strong>Siamese convolutional Neural Networks</strong> for feature extraction.
0 commit comments