Please use this identifier to cite or link to this item:
標題: 以圖騰為索引之影像資料庫設計
Image Database Design Based on Indexing Iconic Pictures
作者: 李朱慧
Lee, Chu-Hui
關鍵字: Image Database;影像資料庫;Spatial Relationship;Signature File;Iconic Picture;Similarity Retrieval;CBIR;Spatial Reasoning;9D-SPA;空間關係;簽章檔;圖騰影像;類似圖搜尋;以內容特徵搜尋圖形;空間推論;九個方位空間表式法
出版社: 應用數學系
在設計影像資料庫系統時,圖片中物件間的空間關係是很重要的特徵,在本篇論文中,我們提出一個稱之為9D-SPA (9-Direction SPanning Area) 以空間關係為主的影像表示法,這個表示法,可以很有效率的實現智慧型影像資料庫的主要功能,包括:圖形重建 (Reconstruction) /瀏覽 (Browsing)、空間推理 (Spatial Reasoning)、圖騰索引 (Iconic Indexing)、及類似圖搜尋 (Similarity Retrieval) 等。若與最小框住物件之方形法 (Minimum Bounding Rectangles) 或是以物件的重心 (Centroids) 為基礎的影像表示法做比較,9D-SPA的識別能力是最有效的,且9D-SPA 表示法在類似圖搜尋功能中可提供模糊比對以符合使用者不同程度的需求。實驗結果顯示本論文提出的搜尋方法有很高的精確度 (Precision) 和召回度 (Recall),此外,9D-SPA 表示法還可以用來建立兩層式的索引 (Two-level Index),以大幅減少搜尋的範圍。論文中也以實驗的方法證實,透過所建立的兩層式索引可以有效率的搜尋圖片。以50,000張圖片的影像資料庫為例,每次搜尋 (Query) 平均大約只需尋找資料庫中0.1254% 到 1.6829% 的影像數量即可 (依不同程度的類似度而有些微不同)。
為了加速影像的搜尋效率,本論文也提出一個影像特徵簽章的索引結構,稱之為 Signature Relation Graph。簽章檔可以快速地過濾大部份不合格的影像,實驗證明,在準確度相同的情況下,Signature Relation Graph在搜尋時比 Quick Filter 的搜尋範圍小,與Signature Tree 的索引法做比較,本論文所提的方法所需比較的位元數也較少。

Spatial relationships between iconic objects are important features for designing a content-based image retrieval system. In this dissertation, we propose a new scheme, called 9D-SPA representation, for encoding the spatial relations in an image. With this representation, important functions of intelligent image database systems such as visualization, browsing, spatial reasoning, iconic indexing, and similarity retrieval can be easily achieved. The capability of discriminating images based on 9D-SPA representation is much more powerful than any spatial representation method based on Minimum Bounding Rectangles or centroids of objects. The similarity measures using 9D-SPA representation provide a wide range of fuzzy matching capability in similarity retrieval to meet different user''s requirements. Experimental results showed that our system is very effective in terms of recall and precision. In addition, the 9D-SPA representation can be incorporated into a two-level index structure to help reducing the search space of each query processing.
To further speed-up the process of image retrieval, we propose a novel structure, called Signature Relation Graph, for organizing signatures. A signature file serves as a search filter to prune off unqualified images at the early stage of image search. The SR graph is superior to the quick filter method and the signature tree method in terms of the number of signatures accessed per query and the number of bits examined per query, respectively.
Based on the 9D-SPA representation, the two-level index structure, and the SR graph signature organization, an overall architecture is proposed in this dissertation for designing an image database system.
Appears in Collections:應用數學系所

Show full item record

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.