Share Email Print

Proceedings Paper

3D image restoration for confocal microscopy: toward a wavelet deconvolution for the study of complex biological structures
Author(s): Jacques Boutet de Monvel; Sophie Le Calvez; Mats Ulfendahl
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Image restoration algorithms provide efficient tools for recovering part of the information lost in the imaging process of a microscope. We describe recent progress in the application of deconvolution to confocal microscopy. The point spread function of a Biorad-MRC1024 confocal microscope was measured under various imaging conditions, and used to process 3D-confocal images acquired in an intact preparation of the inner ear developed at Karolinska Institutet. Using these experiments we investigate the application of denoising methods based on wavelet analysis as a natural regularization of the deconvolution process. Within the Bayesian approach to image restoration, we compare wavelet denoising with the use of a maximum entropy constraint as another natural regularization method. Numerical experiments performed with test images show a clear advantage of the wavelet denoising approach, allowing to `cool down' the image with respect to the signal, while suppressing much of the fine-scale artifacts appearing during deconvolution due to the presence of noise, incomplete knowledge of the point spread function, or undersampling problems. We further describe a natural development of this approach, which consists of performing the Bayesian inference directly in the wavelet domain.

Paper Details

Date Published: 2 May 2000
PDF: 11 pages
Proc. SPIE 3919, Three-Dimensional and Multidimensional Microscopy: Image Acquisition Processing VII, (2 May 2000); doi: 10.1117/12.384187
Show Author Affiliations
Jacques Boutet de Monvel, Karolinska Hospital (Sweden)
Sophie Le Calvez, Karolinska Hospital (Sweden)
Mats Ulfendahl, Karolinska Hospital (Sweden)

Published in SPIE Proceedings Vol. 3919:
Three-Dimensional and Multidimensional Microscopy: Image Acquisition Processing VII
Jose-Angel Conchello; Carol J. Cogswell; Andrew G. Tescher; Tony Wilson, Editor(s)

© SPIE. Terms of Use
Back to Top