National and Kapodistrian University of Athens
Μέθοδοι γραφικών και τεχνητής όρασης για την ανακατασκευή, αναπαράσταση και ανάκτηση τρισδιάστατων αντικειμένων με εφαρμογή στην βιομετρία
Abstract
dc:descriptionThe increase of availability of three dimensional objects, widens their use by computer graphics, computer vision and biometrics applications. Three dimensional data increase the available information about the object, but in the same times introduces new challenges. On most applications where three dimensional object databases are utilized, the goal is their retrieval, which is the ability to compare and categorize these objects. The retrieval procedure requires the optimal reconstruction of the objects as well as the efficient representation of the three dimensional information. In this dissertation, the general problem of managing three dimensional information is tackled, for the reconstruction, the representation and the retrieval of three dimensional objects. Initially, a new method for the reconstruction of three dimensional objects based on Integral Photography is presented. This method extracts the three dimensional information of real-life objects from a grid of elemental images. Additionally, two three dimensional object representations are presented that are utilized by two novel retrieval methods. The reduction of dimensionality is the common feature of both representations. The first representation, based on depth images, is suitable for inter-class object retrieval while the second, based on geometry images, is suitable for intra-class object retrieval. The geometry images are created using a subdivision-based deformable model framework. The intra-class object retrieval method is further specialized and applied in the field of biometrics, since the human body’s geometry is a strong biometric. More specifically, three dimensional information is used to perform face and ear recognition. The performance of the proposed method is evaluated using the largest publicly available biometric databases. The face recognition results that are presented are considered state-of-the-art worldwide.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2008
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Passalis, Georgios
- Πασσαλής, Γεώργιος
Subjects
dc:subject × 20- 3Δ ανακατασκευή
- Παραμορφώσιμα μοντέλα
- Γεωμετρικές εικόνες
- Επιφάνειες υποδιαίρεσης
- Εικόνες βάθους
- Αναγνώριση προσώπου και αυτιού
- Βιομετρία
- Τεχνητή όραση
- 3d reconstruction
- Deformable models
- Geometry images
- Subdivision surfaces
- Depth images
- Face and ear recognition
- Biometrics
- Computer vision
- 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/22590