National Technical University of Athens (NTUA)
Εξαγωγή γεωμετρικής πληροφορίας από απεικονίσεις με φθορά και θόρυβο: ειδικές εφαρμογές στην αρχαιολογία και την ιατρική
Abstract
dc:descriptionThis PhD thesis is divided into two parts.In the first part, the further development of an automated fragmented object reconstruction algorithm is presented. This algorithm uses information extracted from the outline of the fragmented object pieces and was developed by Papaodysseus C., Panagopoulos Th., Exarhos M., Τriantafillou C., Fragoulis D., Doumas C. and presented in the international journal IEEE Trans. On Signal Processing το 2002 (IEEE Trans. On Signal Processing, 50(6), 1277-1288, 2002. The further development of this algorithm involves the potential use of the chromatic and the thematic content of the fragments that constitute the object, in order to confirm or reject the proposals for potential matches that are the result of the application of the initial form of the fragmented object reconstruction algorithm. In the initial version of the reconstruction algorithm, only the information extracted from the fragment outline is used. The improved algorithm was initially applied in the reconstruction of a simple commercial jigsaw puzzle, in order to ascertain the effectiveness of the algorithm in excellent quality images and in fragments that have not been degraded due to wear and the passage of time. Then, the improved algorithm was applied in the real-life case of the reconstruction of fragmented wall-paintings that were unearthed during the archaeological excavation of the Akrotiri site in Santorini.In the second part of the thesis, the application of image segmentation techniques in radiology images (which are low quality images) is presented, in order to accurately determine the depicted projection of a special nail, which is constructed out of a special slightly flexible titanium alloy, and is used in closed intramedullary nailing surgical procedures. These particular procedures are used in treating long fragmented or broken bones (hip bone, tibia fractures, etc). Using shape and pattern recognition techniques, the maximum geometric information from the x-ray projection of these nails was extracted. This information will be then used for the exact determination of the position and the shape of the nail, as well as the orientation of the nail with regard to the fragmented bone, in the interior of which the nail is surgically inserted. Thus, the fixation of the nail with the use of screws is accomplished and the surgical procedure is completed, even if the nail has been slightly deformed during its insertion in the bone interior.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2013
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Skembris, Aggelos-Saverios
- Σκέμπρης, Άγγελος-Σαβέριος
Subjects
dc:subject × 20- Αναγνώριση προτύπων
- Ανακατασκευή θραυσμένων αντικειμένων
- Χρωματικό περιεχόμενο
- Θεματικό περιεχόμενο
- Ακτινογραφία
- Κατάτμηση εικόνας
- Αναγνώριση σχήματος
- Κλειστή ενδομυελική ήλωση
- Pattern recognition
- Fragmented object reconstruction
- Chromatic content
- Thematic content
- Radiology images
- Image segmentation
- Shape recognition
- Closed intramedullary nailing
- 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/28659