Direct edge-preserving regularization in tomographic image reconstruction

ePubs

The open archive for STFC research publications

Persistent URL	http://purl.org/net/epubs/work/12317845
Record Status	Checked
Record Id	12317845
Title	Direct edge-preserving regularization in tomographic image reconstruction
Contributors	D Kazantsev, E Ovtchinnikov (STFC Rutherford Appleton Lab.), WRB Lionheart, PJ Withers, PD Lee
Abstract	In this paper, we present a novel regularization technique for iterative tomographic reconstruction. The proposed penalty is built on the edge-preserving Laplacian and encourages piecewise-smooth solutions with sharp edges. Our penalty outperforms total variation regularization, resulting in higher signal-to-noise ratio and visually more appealing reconstructed images. Simple and effective implementation of the proposed regularization technique within the conjugate gradient algorithm makes it a competitive alternative to conventional regularization methods.
Organisation	STFC , SCI-COMP , SCI-COMP-CM
Keywords
Funding Information
Related Research Object(s):
Licence Information:
Language	English (EN)

Type	Details	URI(s)	Local file(s)	Year
Preprint	RAL-P-2015-001, Proceedings of the Royal Society A 2015.		RAL-P-2015-001.pdf	2015