Conference proceeding
Improved higher degree total variation (HDTV) regularization
2012 9th IEEE International Symposium on Biomedical Imaging (ISBI), pp.656-659
05/2012
DOI: 10.1109/ISBI.2012.6235633
Abstract
The main focus of this paper is to further improve the performance of the recently introduced higher degree total variation (HDTV) penalties, which are L 1 -L p ; p ≥ 1 norms of directional image derivatives. We generalize this class as the L 1 -L p norms of image responses to rotated versions of an arbitrary derivative operator. We show that several penalties proposed by other researchers are special cases of the generalized isotropic penalties (p = 2), when the derivative operator is chosen appropriately; our experiments show that the anisotropic (p = 1) versions of these penalties provide improved reconstructions. In addition, we optimize the derivative operator for improved orientation selectivity, thus further improving the ability of the resulting penalties to provide high quality image reconstructions. We also focus on the efficient discretization of HDTV penalties, which are specified in the continuous domain. Specifically, we approximate the derivative operators as the sum of partial derivatives of an almost isotropic B-spline window. Our numerical experiments confirm the benefit of the improved discretization and the optimization of the operator.
Details
- Title: Subtitle
- Improved higher degree total variation (HDTV) regularization
- Creators
- Yue Hu - Dept. of Electr. & Comput. Eng., Univ. of Rochester, Rochester, NY, USAM Jacob - Dept. of Electr. & Comput. Eng., Univ. of Iowa, Iowa City, IA, USA
- Resource Type
- Conference proceeding
- Publication Details
- 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI), pp.656-659
- Publisher
- IEEE
- DOI
- 10.1109/ISBI.2012.6235633
- ISSN
- 1945-7928
- eISSN
- 1945-8452
- Language
- English
- Date published
- 05/2012
- Academic Unit
- Electrical and Computer Engineering; Roy J. Carver Department of Biomedical Engineering; Radiology; Iowa Neuroscience Institute; Radiation Oncology
- Record Identifier
- 9984070686002771
Metrics
5 Record Views