Browsing by Author "Mequanint, Dessalegn"

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    SIMILARITY-BASED VIDEO RETRIEVAL: MODELING and PROCESSING
    (Addis Ababa University, 2004-06) Mequanint, Dessalegn; Atnafu, Solomon(PhD)
    Nowadays, digital video data is increasingly available for public accesses. However, the rate at which access to this key element of multimedia computing is growing is unparalleled with the tools and techniques that have been developed for retrieving and browsing it. In recent years, a lot of research efforts have been put into video data management, amongst other active research topics, video data modeling has been extensively researched, but nearly no work has been done in the area of video data modeling that reveals the role of metadata in enhancing video retrieval systems, and only little work has been done in the area of video data repository modeling. In many cases the motivation of the research efforts has been the accessibility problem. In this thesis, a video data model with two separate representation schemes is proposed. The proposed video data model distinguishes three parts: the frame-based scheme, the object-based scheme and the external description block. We focus our attention on the frame-based scheme and the external description block. In addition, a video data repository model that can be conveniently used under OR-DBMSs environment is proposed. On the basis of the proposed video data repository models, a similarity-based video retrieval technique in the context of ORDBMS is also proposed. In an effort towards the design of effective and interoperable video retrieval systems, standards play a major role, in this thesis through the video data model we proposed the benefits of MPEG- 7 are directed to reach OR-DBMSs. Furthermore, to demonstrate the practicality of our proposals a prototype system for a soccer application domain which is called SVAMS is developed. Key words: Video data model, Video data repository model, Frame-based representation, Object-based representation, Video databases, Similarity-based video retrieval
  • No Thumbnail Available
    Item
    SIMILARITY-BASED VIDEO RETRIEVAL: MODELING and PROCESSING
    (Addis Ababa University, 2004-06) Mequanint, Dessalegn; Atnafu, Solomon(PhD)
    Nowadays, digital video data is increasingly available for public accesses. However, the rate at which access to this key element of multimedia computing is growing is unparalleled with the tools and techniques that have been developed for retrieving and browsing it. In recent years, a lot of research efforts have been put into video data management, amongst other active research topics, video data modeling has been extensively researched, but nearly no work has been done in the area of video data modeling that reveals the role of metadata in enhancing video retrieval systems, and only little work has been done in the area of video data repository modeling. In many cases the motivation of the research efforts has been the accessibility problem. In this thesis, a video data model with two separate representation schemes is proposed. The proposed video data model distinguishes three parts: the frame-based scheme, the object-based scheme and the external description block. We focus our attention on the frame-based scheme and the external description block. In addition, a video data repository model that can be conveniently used under OR-DBMSs environment is proposed. On the basis of the proposed video data repository models, a similarity-based video retrieval technique in the context of ORDBMS is also proposed. In an effort towards the design of effective and interoperable video retrieval systems, standards play a major role, in this thesis through the video data model we proposed the benefits of MPEG- 7 are directed to reach OR-DBMSs. Furthermore, to demonstrate the practicality of our proposals a prototype system for a soccer application domain which is called SVAMS is developed. Key words: Video data model, Video data repository model, Frame-based representation, Object-based representation, Video databases, Similarity-based video retrieval