Design of Iconic Indexing Strategies for Multimedia Database Systems
| Autor: | Ann Hsing-Yen, 安興彥 |
|---|---|
| Rok vydání: | 1998 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 86 One of the most important problems in the design of image database systems is how images are stored in the image database. In traditional database systems, the use of indexing to allow database accessing has been well established. Analogously, picture indexing techniques are needed to make ease pictorial information retrieval from a pictorial database.In the previous approaches to represent pictorial data, as thecomplexity of the representation strategy is increased, the morespatial relationships can be represented, which also results in a more complex strategy for query processing and a limited types of queries that can be answered. Therefore, in this thesis, first, we propose an efficient iconic indexing scheme called Unique-ID-based Matrix (UID Matrix) for symbolic pictures, which combines the advantages of the 2D C-string and the 9DLT matrix. In the proposed scheme, we carefully assign each spatial operator a unique identifier such that the processing of category classification becomes the range checking operation; therefore, they are efficient enough as compared to the previous approaches. Second, in the previous approaches to retrieve a symbolic picture from an image database which contains multiple symbolic pictures, they only consider pictures of type-2 similarity, where a picture A is of type-2 similarity with picture B when both pictures have the same spatial relationship between any two objects in the pictures. When queries of objects, type-0 similarity and type-1 similarity are asked, signatures for objects, type-0 similarity and type-1 similarity are required. Although these 4 kinds of signatures can be constructed and stored in advance to speed up answering such kinds of queries as in Lee et al.'s approach, such a large signature file containing those signatures really waste space. In this thesis, we first present revised versions of Chang and Yang's two access methods, which construct type-2 signatures. Based on these revised versions, we then propose algorithms to convert those type-2 signatures into object signatures, type-0 signatures and type-1 signatures so that we can answer queries of object similarity, type-0 similarity, type-1 similarity and type-2 similarity by storing only one type of signatures, i.e., the type-2 signatures. While in Lee et al.'s approach, they have to construct 4 types of signatures in the database to achieve the same goal. Moreover, we will show that our approach can provide a higher flexibility in terms of storage space and a lower rate of false match than Lee et al.'s approach. Third, the adaptive 2D-H string representation has been proposed to remove the redundancy existing in the 2D-H string representation. However, the concise representation of the adaptive 2D-H string can cause ambiguous cases. A picture is defined to be ambiguous if there exists more than one different reconstructed picture from its representation. In this thesis, we show such a case and provide an answer to avoid the ambiguous case. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|