Abstract
dc:descriptionCoping with the limited spatial resolution of images is the objective of this thesis. Initially, an effort to model the scanner function by means of simple models is made. Afterwards, a nonuniform interpolation method for Super-Resolution (SR) image by means of Kriging interpolation is proposed. Experimentation proves that the proposed method predominates over method which employs the weighted nearest neighbor interpolation. Also, three new methods for stochastic regularized SR image reconstruction are presented. The Tukey, Lorentzian and Huber estimators in combination with the bilateral total variation regularization are employed. Experimental results verify the robust statistics theory. Moreover, a novel study the conclusions of which help to select and effective SR image reconstruction method for a given low-resolution sequence of frames is carried out. Finally, an image interpolation method employing a neural network is proposed.
Degree
thesis:*- Grantor dc:publisher
- University of Patras
- Year dc:date
- 2009
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Panagiotopoulou, Antigoni
- Παναγιωτοπούλου, Αντιγόνη
Subjects
dc:subject × 23- Μοντελοποίηση λειτουργίας ψηφιοποιητή εικόνας
- Βελτίωση χωρικής ανάλυσης
- Ανακατασκευή εικόνας super-resolution
- Παρεμβολή Kriging
- Εκτιμητές Tukey
- Επιλογή αποτελεσματικής μεθόδου super-resolution
- Παρεμβολή μέσω νευρωνικού δικτύου
- Εκτιμητές Lorentzian
- Εκτιμητές Huber
- Scanner function modeling
- Spatial resolution improvement
- Super-resolution image reconstruction
- Kriging interpolation
- Tukey estimators
- Bilateral total variation regularization
- Selection of effective super-resolution method
- Interpolation by means of neural network
- Lorentzian estimators
- Huber estimators
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Natural Sciences
- Computer and Information Sciences
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/26614