Sparse image extrapolation using different inpainting algorithms
The result's identifiers
Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26220%2F12%3APU100049" target="_blank" >RIV/00216305:26220/12:PU100049 - isvavai.cz</a>
Result on the web
—
DOI - Digital Object Identifier
—
Alternative languages
Result language
angličtina
Original language name
Sparse image extrapolation using different inpainting algorithms
Original language description
Image inpainting via approximately solving underdetermined systems of linear equations can take different forms. State of the art methods use sparse solutions of such systems to inpaint (i.e. fill-in) the missing part of an image. Some of these approaches are applicable for image extrapolation as well, but this cannot be seen just as a special case of standard inpainting problem. For example, usual methods assume filling the holes from different directions, which is not tractable in the case of extrapolation. In this paper some of the algorithms that are tailored to inpainting are introduced and modified for use in image extrapolation. We use K-SVD algorithm that trains a dictionary for optimal sparse representation, MCA (Morphological Component Analysis) that expects two incoherent dictionaries for representing separately cartoon and texture. The last algorithm present is the statistics-based EM (Expectation Maximization). The performance of these algorithms for image extrapolation is
Czech name
—
Czech description
—
Classification
Type
D - Article in proceedings
CEP classification
JA - Electronics and optoelectronics
OECD FORD branch
—
Result continuities
Project
—
Continuities
S - Specificky vyzkum na vysokych skolach
Others
Publication year
2012
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data specific for result type
Article name in the collection
Proceedings of the 14th International Conference on Research in Telecommunication Technologies
ISBN
978-80-554-0569-8
ISSN
—
e-ISSN
—
Number of pages
5
Pages from-to
247-251
Publisher name
Neuveden
Place of publication
Neuveden
Event location
Vrátna, SK
Event date
Sep 1, 2012
Type of event by nationality
EUR - Evropská akce
UT code for WoS article
—