Restauração de imagens utilizando projeções em conjuntos convexos e algoritmos evolucionistas

• Ano de defesa: 2014
• Autor(a) principal: Pires, Rafael Gonçalves [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Computação; Processamento de imagens - Tecnicas digitais; Reconstrução de imagens; Conjuntos convexos; Otimização matematica; Image processing - Digital techniques
• Link de acesso: http://hdl.handle.net/11449/110378

The process of image restoration aims to enhance images corrupted by noise and blurred. Iterative techniques can better control the restoration algorithm in order to restore blurred regions in details without increasing noise. Techniques based on Projection Sets in Convex (Projections onto Convex Sets - POCS) have been used in the context of image restoration by projecting the solution in a hyperspace until some convergence criterion is met. The expected result is a better picture at the end of an unknown number of projections. The number of convex sets and its combinations allow you to build several image restoration algorithms based on POCS. This study uses two convex sets: Row Action Projections (RAP) and Limited Amplitude (LA). The RAP algorithm has a relaxation parameter depends on the characteristics of the image that will be restored. Thus, erroneous values of can lead to a poor restoration. We propose to find the value of as the problem of modeling and optimization using different evolutionary techniques. Furthermore, is possible to use the parameters learned in restoring an image, and use them to another image