|
This article is cited in 1 scientific paper (total in 1 paper)
INFORMATICS
Suppression of speckle noise in medical images via segmentation-grouping of 3D objects using sparse contourlet representation
V. F. Kravchenkoab, Yu. V. Gulyaeva, V. I. Ponomarevc, G. Aranda Bojorgesc a Kotelnikov Institute of Radioengineering and Electronics, Russian Academy of Sciences, Moscow, Russia
b Bauman Moscow State Technical University, Moscow, Russia
c Instituto Politecnico Nacional de Mexico, Mexico City, Mexico
Abstract:
A novel method for filtering ultrasonic and magnetic resonance images contaminated by multiplicative (speckle) noise is justified and implemented. The method consists of several stages: segmentation of image areas, grouping of similar structures in 3D, homomorphic transformation, a 3D filtering approach based on a sparse representation in the contourlet transform (CLT) space with posterior filtering according to MI weights of similar 2D structures, and the final inverse homomorphic transformation. A physical interpretation of the filtering procedure in the case of speckle noise is given, and a structural scheme for noise suppression is developed. Simulation based on the new filtering procedure has confirmed its superiority in terms of generally accepted criteria, such as the structural similarity index measure, peak signal-to-noise ratio, edge preservation index, and the resolution index alpha, as well as in a visual comparison of filtered images.
Keywords:
ultrasonic and magnetic resonance images, superpixel segmentation methods, filtering, speckle noise, grouping of objects, holomorphic transformation, peak signal-to-noise ratio.
Received: 09.09.2022 Revised: 24.11.2022 Accepted: 26.12.2022
Citation:
V. F. Kravchenko, Yu. V. Gulyaev, V. I. Ponomarev, G. Aranda Bojorges, “Suppression of speckle noise in medical images via segmentation-grouping of 3D objects using sparse contourlet representation”, Dokl. RAN. Math. Inf. Proc. Upr., 509 (2023), 94–100; Dokl. Math., 107:1 (2023), 77–82
Linking options:
https://www.mathnet.ru/eng/danma368 https://www.mathnet.ru/eng/danma/v509/p94
|
Statistics & downloads: |
Abstract page: | 73 | References: | 29 |
|