Extracting textual information from images and videos for automatic content-based annotation and retrieval [Elektronische Ressource] / vorgelegt von Julinda Gllavata

Extracting Textual Information from Imagesand Videos for Automatic Content-BasedAnnotation and RetrievalDissertationzur Erlangung des akademischen GradesDoktor der Naturwissenschaften (Dr. rer. nat)demFachbereich Mathematik und Informatik derPhilipps-Universit¨at Marburgvorgelegt vonJulinda Gllavataaus Durr¨es, AlbanienMarburg/Lahn, 2007Vom Fachbereich Mathematik und Informatik derPhilipps-Universit¨at Marburg als Dissertation am29. Januar 2007 angenommen.Erstgutachter: Prof. Dr. Bernd Freisleben, Philipps-Universit¨at MarburgZweitgutachter: Prof.Dr.ManfredSommer,Philipps-Universit¨atMarburgTag der mu¨ndlichen Pru¨fung: 5. Februar 2007DedicationTo my lovely parents Esi and Meti.Prind¨erve t¨e mi t¨e dashur Esit dhe Metit.AcknowledgementsFirst of all, I would like to express my gratitude to my advisor Prof. Dr.Bernd Freisleben for his invaluable support, guidance and encouragementthroughout this research project. Without the helpful discussions, sugges-tions and insights, this work would not have been completed.I would like to thank Prof. Dr. Manfred Sommer for accepting to act asthe second reviewer of this thesis.The research work presented in this thesis was financially supported bythe Deutsche Forschungsgemeinschaft (SFB/FK 615, Teilprojekt MT) andby Deutscher Akademischer Austausch Dienst (DAAD, Stability Pact forSouth Eastern Europe). The support of these two institutions is gratefullyacknowledged. In this context, I wish to thank Prof.