by Wei, T and Klette, R
Abstract:
This paper presents a wavelet-based algorithm for height from gradients. The tensor product of the third-order Daubechies’ scaling functions is used to span the solution space. The surface height is described as a linear combination of a set of the scaling basis functions. This method efficiently discretizes the cost function associated with the height from gradients problem. After discretization, the height from gradients problem becomes a discrete minimization problem rather than discretized PDE’s. To solve the minimization problem, perturbation method is used. The surface height is finally decided after finding the weight coefficients.
Reference:
A wavelet-based algorithm for height from gradients (Wei, T and Klette, R), In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Springer Verlag, volume 1998, 2001.
Bibtex Entry:
@inproceedings{wei2001agradients,
author = "Wei, T and Klette, R",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "84--90",
publisher = "Springer Verlag",
title = "A wavelet-based algorithm for height from gradients",
volume = "1998",
year = "2001",
abstract = "This paper presents a wavelet-based algorithm for height from gradients. The tensor product of the third-order Daubechies’ scaling functions is used to span the solution space. The surface height is described as a linear combination of a set of the scaling basis functions. This method efficiently discretizes the cost function associated with the height from gradients problem. After discretization, the height from gradients problem becomes a discrete minimization problem rather than discretized PDE’s. To solve the minimization problem, perturbation method is used. The surface height is finally decided after finding the weight coefficients.",
isbn = "3540416943",
issn = "0302-9743",
eissn = "1611-3349",
language = "eng",
}