ML Anthology

Robust Fusion of Dynamic Shape and Normal Capture for High-Quality Reconstruction of Time-Varying Geometry

Naveed Ahmed, Christian Theobalt, Petar Dobrev, Hans-Peter Seidel, Sebastian Thrun
CVPR 2008
doi:10.1109/CVPR.2008.4587696 /cvpr/2008/ahmed2008cvpr-robust/

Abstract

We present a new passive approach to capture time-varying scene geometry in \nlarge acquisition volumes from multi-view video. It can be applied to \nreconstruct complete moving models of human actors that feature even slightest \ndynamic geometry detail, such as wrinkles and folds in clothing, and that can \nbe viewed from 360 degrees. Starting from multi-view video streams recorded \nunder calibrated lighting, we first perform marker-less human motion capture \nbased on a smooth template with no high-frequency surface detail. Subsequently, \nsurface reflectance and time-varying normal fields are estimated based on the \ncoarse template shape. The main contribution of this work is a new statistical \napproach to solve the non-trivial problem of transforming the captured normal \nfield that is defined over the smooth non-planar 3D template into true 3D \ndisplacements. Our spatio-temporal reconstruction method outputs displaced \ngeometry that is accurate at each time step of video and temporally smooth, \neven if the input data are affected by noise.

CVPR Semantic Scholar

Cite

Text

Ahmed et al. "Robust Fusion of Dynamic Shape and Normal Capture for High-Quality Reconstruction of Time-Varying Geometry." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2008. doi:10.1109/CVPR.2008.4587696

Markdown

[Ahmed et al. "Robust Fusion of Dynamic Shape and Normal Capture for High-Quality Reconstruction of Time-Varying Geometry." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2008.](https://mlanthology.org/cvpr/2008/ahmed2008cvpr-robust/) doi:10.1109/CVPR.2008.4587696

BibTeX

@inproceedings{ahmed2008cvpr-robust,
  title     = {{Robust Fusion of Dynamic Shape and Normal Capture for High-Quality Reconstruction of Time-Varying Geometry}},
  author    = {Ahmed, Naveed and Theobalt, Christian and Dobrev, Petar and Seidel, Hans-Peter and Thrun, Sebastian},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2008},
  doi       = {10.1109/CVPR.2008.4587696},
  url       = {https://mlanthology.org/cvpr/2008/ahmed2008cvpr-robust/}
}