21 lines
1.4 KiB
TeX
21 lines
1.4 KiB
TeX
|
\chapter*{Introduction}
|
||
|
|
|
||
|
|
As a conclusion to our studies at the ENSEEIHT, we have been working on our Projet Long. Our project takes part in Bastien Durix thesis which focuses on the extraction of 2D skeletons from pictures and the computation of 3D skeletons from associated 2D skeletons of the same objects, and will also help him on other parts of his approach.
|
||
|
|
|
||
|
|
Our subject consisted in realizing animated 3D versions of real 3D objects. We based ourselves on different pictures of objects taken from different perspectives, and use them to reconstruct a 3D version of the object, which we animate.\\
|
||
|
|
|
||
|
|
Here is our pipeline :
|
||
|
|
|
||
|
|
\begin{itemize}
|
||
|
|
\item segmentation
|
||
|
|
\item intrinsic and extrinsic camera calibration
|
||
|
|
\item points detection and matching
|
||
|
|
\item skeletonization : cut and match skeleton pieces
|
||
|
|
\item computing the B-splines
|
||
|
|
\item meshing the 3D skeleton computed with the 2D skeletons
|
||
|
|
\item meshing of extremities and junctions
|
||
|
|
\item choosing connecting points to animate the object
|
||
|
|
\item animation
|
||
|
|
\end{itemize}
|
||
|
|
|
||
|
|
We were two Projet Long groups working on these subject. The first part was split into two subsections : the other group got to work on the segmentation and B-splines computation, while we were given the matching part and the camera calibration. The second part, concerning the meshing and the animation, was done completely by the two groups.
|