Method of prediction based on algorithmic probability in the problem of image restoration in missing regions

Autor: O. V. Shcherbakov, Alexey Potapov, Innokentii Zhdanov
Rok vydání: 2013
Předmět:
Zdroj: Journal of Optical Technology. 80:680
ISSN: 1070-9762
Popis: This paper discusses the criterion of algorithmic probability, which offers a general solution to the question of extrapolating symbolic strings. The given criterion extends the theoretical-informational approach based on algorithmic complexity, which is widely used to synthesize image-analysis methods. Methodological recommendations are given concerning the practical application of the algorithmic-probability criterion when processing and analyzing images. As an example, a method is developed for restoring images in missing regions, which can be implemented on a computer by narrowing an algorithmically complete space of models of images to a set of their Fourier transforms.
Databáze: OpenAIRE