Significance of Elastic Properties in Physics-Based Nonrigid Motion Modeling, a Numerical Sensitivity Analysis

Abstract

Parameters used in the physical models for nonrigid motion simulation are often not known with high precision. It has been recognized that the commonly used assumptions about the parameters may have adverse effects on the modeling quality. We present an efficient sensitivity analysis method to assess the impact of those assumptions by examining the model’s response to the parameter perturbation. Numerical experiments with both synthetic and real models show that (1) the normalized sensitivity distribution can help determine the relative importance of different parameters; (2) the dimensional sensitivity is useful in the assessment of a particular parameter assumption; (3) the model is more sensitive at the property discontinuity (heterogeneity). The proposed sensitivity analysis method is general and can also be applied to the assessment of other types of assumptions such as nonlinearity and anisotropy.

Cite

Text

Zhang et al. "Significance of Elastic Properties in Physics-Based Nonrigid Motion Modeling, a Numerical Sensitivity Analysis." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2004. doi:10.1109/CVPR.2004.437

Markdown

[Zhang et al. "Significance of Elastic Properties in Physics-Based Nonrigid Motion Modeling, a Numerical Sensitivity Analysis." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2004.](https://mlanthology.org/cvpr/2004/zhang2004cvpr-significance/) doi:10.1109/CVPR.2004.437

BibTeX

@inproceedings{zhang2004cvpr-significance,
  title     = {{Significance of Elastic Properties in Physics-Based Nonrigid Motion Modeling, a Numerical Sensitivity Analysis}},
  author    = {Zhang, Yong and Goldgof, Dmitry B. and Sarkar, Sudeep},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2004},
  pages     = {21},
  doi       = {10.1109/CVPR.2004.437},
  url       = {https://mlanthology.org/cvpr/2004/zhang2004cvpr-significance/}
}