|
This article is cited in 1 scientific paper (total in 1 paper)
INFORMATICS
Suppression of multiplicative noise in images via grouping of similar objects
V. F. Kravchenkoabc, V. I. Ponomarevd, V. I. Pustovoĭtbc, G. Aranda Bojorgesd a Kotel'nikov Institute of Radio Engineering and Electronics, Russian Academy of Sciences, Moscow, Russia
b Scientific and Technological Centre of Unique Instrumentation, Russian Academy of Sciences, Moscow, Russia
c Bauman Moscow State Technical University, Moscow, Russia
d Instituto Politecnico Nacional de Mexico, Mexico City, Mexico
Abstract:
For the first time, a method for filtering of images distorted by multiplicative noise has been substantiated and implemented. The method proposed consists of several stages: formation of 3D similar structures, homomorphic transformation, image filtering in the 3D space of discrete cosine transform (DCT), MAP estimation of images via grouping of similar structures, inverse homomorphic transformation, and the final stage of processing, in which the errors are corrected and edges and details of the images are recovered. Physical interpretation of the filtering procedure under conditions of multiplicative noise is given, and structural scheme of noise suppression is developed. Simulation of the method proposed has confirmed the advantages of the new filtering procedure in terms of the generally recognized criteria: estimation of structural similarity index measure and peak signal-to-noise ratio, and edge preservation index, as well as in visual comparison of filtered images.
Keywords:
image, filtering, multiplicative noise, grouping of objects, homomorphic transformation, peak signal-to-noise ratio.
Received: 28.04.2021 Revised: 16.05.2021 Accepted: 26.05.2021
Citation:
V. F. Kravchenko, V. I. Ponomarev, V. I. Pustovoǐt, G. Aranda Bojorges, “Suppression of multiplicative noise in images via grouping of similar objects”, Dokl. RAN. Math. Inf. Proc. Upr., 499 (2021), 67–72; Dokl. Math., 104:1 (2021), 216–220
Linking options:
https://www.mathnet.ru/eng/danma193 https://www.mathnet.ru/eng/danma/v499/p67
|
Statistics & downloads: |
Abstract page: | 73 | Full-text PDF : | 16 | References: | 13 |
|