Title: Wavelet-based compression/denoising of coronary X-ray images

Authors: Azza Ouled Zaid, Ammar Bouallegue, Christian Olivier, Amine Nait-Ali

Addresses: SysCom Laboratory, National Engineering School of Tunis, BP 37, 1002 le Belvedere, Tunis, Tunisia. ' SysCom Laboratory, National Engineering School of Tunis, BP 37, 1002 le Belvedere, Tunis, Tunisia. ' XLIM Laboratory, UMR CNRS 6172, SIC (Signal, Images et Communications) Department, Poitiers University, Bat. SP2MI, Teleport 2, Boulevard M. et P. Curie, F-86962 Futuroscope-Chasseneuil cedex, France. ' LISSI Laboratory, Paris 12 University, Creteil, France

Abstract: Compression of angiographic images has been shown to be difficult when compared with other medical imaging modalities. The factors partially responsible for this are the presence of complex structures that are only apparent by subtle changes in the contrast and the significant amount of acquisition noise. In this work, we propose a Comp–Denoiser adapted to coronary X-ray images. For this purpose, Wavelet-based Trellis Coded Quantisation (WTCQ) algorithm is extended to incorporate a bivariate thresholding that considers dependencies between wavelet coefficients and their parents in coarser sub-bands. Experimental results show that despite its simplicity our method yields high compression performance.

Keywords: coronary X-ray images; wavelet transform; WTCQ coder; bivariate thresholding; statistic modelling; image compression; image denoising; angiographic images; medical imaging; trellis coded quantisation.

DOI: 10.1504/IJSISE.2009.029324

International Journal of Signal and Imaging Systems Engineering, 2009 Vol.2 No.1/2, pp.7 - 16

Published online: 19 Nov 2009 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article