Estamos realizando la búsqueda. Por favor, espere...


IFS-Based image reconstruction of binary images with functional networks

Abstract: This work addresses the IFS-based image reconstruction problem for binary images. Given a binary image as the input, the goal is to obtain all the parameters of an iterated function system whose attractor approximates the input image accurately; the quality of this approximation is measured according to a similarity function between the original and the reconstructed images. This paper introduces a new method to tackle this issue. The method is based on functional networks, a powerful extension of neural networks that uses functions instead of the scalar weights typically found in standard neural networks. The method relies on an artificial network comprised of several functional networks, one for each of the contractive affine maps forming the IFS. The method is applied to an illustrative and challenging example of a fractal binary image exhibiting a complicated shape. The graphical and numerical results show that the method performs very well and is able to reconstruct the input image using IFS with high accuracy. The results also show that the method is not yet optimal and offers room for further improvement

Otras publicaciones de la misma revista o congreso con autores/as de la Universidad de Cantabria

 Autoría: Gálvez A., Fister I., Iglesias A., Fister I., Gómez-Jauregui V., Manchado C., Otero C.,

 Fuente: Mathematics, 2022, 10, 1107

Editorial: MDPI

 Fecha de publicación: 29/03/2022

Nº de páginas: 26

Tipo de publicación: Artículo de Revista

 DOI: 10.3390/math10071107

ISSN: 2227-7390

 Proyecto español: TIN2017-89275-R

 Proyecto europeo: info:eu-repo/grantAgreement/EC/H2020/778035/EU/PDE-based geometric modelling, image processing, and shape reconstruction/PDE-GIR/