Ampliação de Imagens utilizando Algoritmo de Interpolação Não-Linear, Localmente Adaptativo e Preservador de Bordas

Leízza F. Rodrigues, Rodrigo P. Lemos & Luiz Marcos G. Gonçalves

{leí[email protected],[email protected],[email protected]}

M.Sc. Dissertation Abstract

           The problem of producing an enlarged picture from a given digital image does not possess a definitive solution. Although it has techniques of interpolation capable to create a bigger picture and with certain degree of smoothing, it still has serious difficulties in the handling of details of the picture, as its edges. Generally, interpolation processes tend to mask or to attenuate details in favor of a better smoothing. A central question is how to keep at the same time the details and the smoothing, in a way to add more quality to the picture. This work deals with this problem considering a new interpolation method that is nonlinear, local adaptive, and edge preserving. Thus, the edges of the image are preserved and it is, as a whole, alleviated. The results are compared with common existing interpolation techniques in the literature, applied to different types of images.

Key-words: Resampling, Interpolation.

Acknowledgements

Author L. Rodrigues was supported by a bursary from CAPES – MEC, program DS/PROAP. We would like to thank Dr. Dibio L. Borges from PUC-PR, Curitiba, Brazil, for his valuable contribution and amazing ideas.