Real-time 3D Shape Reconstruction and Refinement from Multi-viewpoint Image Sequences.
|
Accession number;02A0234239
|
| Title;Real-time 3D Shape Reconstruction and Refinement from Multi-viewpoint Image Sequences. |
| Author;
WU X
(Kyoto Univ., Graduate School of Informatics, JPN)
NOBUHARA SHOHEI
(Kyoto Univ., Graduate School of Informatics, JPN)
WADA TOSHIKAZU
(Kyoto Univ., Graduate School of Informatics, JPN)
MATSUYAMA TAKASHI
(Kyoto Univ., Graduate School of Informatics, JPN)
|
Journal Title;Joho Shori Gakkai Kenkyu Hokoku
|
Journal Code:Z0031B
|
ISSN:0919-6072
|
|
VOL.2002;NO.2(CVIM-131);PAGE.61-68(2002)
|
| Figure&Table&Reference;FIG.14, REF.9 |
| Pub. Country;Japan |
| Language;Japanese |
| Abstract;This paper presents two methods for 3D shape reconstruction from multi-viewpoint images. One is volume intersection and the other is shape refinement by deformable mesh model. The volume intersection method is used for initial shape estimation, because of its stability and suitability for whole shape reconstruction. In this method, a parallel volume intersection algorithm and implementation method for PC cluster system are shown. In the next method, the result of volume intersection is converted to the surface mesh model, which dynamically changes its shape according to 1) "photo-consistency", 2)silhouette constraint, 3) and smoothness constraint. Some experimental results demonstrating that this method improves shape accuracy are shown. (author abst.) |
|
|
|
Related Articles;
|
|
