以本體為註記基礎應用於圖像資源管理

對於圖像或影音等的非文字性物件，通常是利用文字描述的註記，以協助擷取及辨識物件內容，傳統上，註記是以事先定義的格式，再由關鍵字檢索方式，找出符合條件之標的物；然而，註記內容缺乏在語意上的概念，而註記格式之間也缺乏彼此的關聯，導致辨識物件內容的效能受限，進而使得檢索效能不佳。本研究利用建立知識本體的方式取代表格式的註記，由於本體方法具有定義知識概念的階層及邏輯特性，因此導入領域知識後可應用於推論機制，使得圖像物件的檢索提升為語意擷取；本研究並採用OWL DL做為知識本體的描述格式，並建立以維管束植物圖像為對象的資訊擷取系統，將可提供於非文字性物件內容辨識上的設計參考，並透過實證經驗上的發現來提供後續研究的修正。

關鍵字

註記；非文字性物件； OWL DL ；知識本體

並列摘要

Technology in the field of digital media generates huge amounts of non-textual objects, audio, video, and images, along with more familiar textual information. Conventional information retrieval employs a keyword-based or even full-text search engine to extract useful information of textual contents. The same techniques are used to exploring non-textual information by creating metadata models and annotations. However, keywords alone lack precision, especially in retrieving items from very large collections, due to the lack of semantic. Keyword-based search lacks the domain knowledge required to pick out relationships between the extracted entities, and the words may not appropriate representations for their meanings and insufficient to retrieve all the relevant information. In addition, users may not know exactly what they looking for and what proper scientific words should give a search engine. This research aims to address semantic issues of information retrieval by developing a knowledge-aware approach that utilizes ontology-based annotation for describing information objects. Ontology is a collection of concepts and their interrelationships, which provide an abstract view of an application domain. A domain ontology may be defined as the specification of a representational vocabulary for a shared domain of discourse which may include definitions of classes, relations, functions and other objects. Non-textual information retrieval is improved by the ability to perform searches, which exploit the ontology to make inferences about data from heterogeneous resources. This research particularly regards a scenario that heavily relies on photo annotations. We develop an ontology to represent the vascular plant domain. The National Museum of Nature Science (NMNS) will provide both vascular plant photo sets and related digital archives as test samples for our research. We also implement an information retrieval system which specialized in recovering vascular plant. In this system, OWL DL is used to be the ontology language. The OWL DL is one of OWL (Web Ontology Language) versions, which performs the concepts of description language to present relationships in a logical form. This research expects to contribute a guidance that can be facilitated to manage digital contents in a knowledge way.

並列關鍵字

Ontology ； Annotations ； Non-textual Objects ； OWL DL

參考文獻

[1] Alani, H., Kim, S., Millard, D.E., Weal, M.J., Hall, W., Lewis, P.H., and Shadbolt, N.R., "Automatic Ontology-Based Knowledge Extraction from Web Documents," IEEE Intelligent Systems, Vol. 18, No. 1, 2003, pp. 14-21.

[2] Baxter, M., Product Design: Practical methods for the Systematic Development of New Products, Chapman and Hall, London, 1995.

[6] Chavez-Aragon, A., and Starostenko, O., "Ontological Shape-Description: A New Method for Visual Information Retrieval," Electronics, Communications and Computers, 2004. CONIELECOMP 2004. 14th International Conference, Feb. 16-18, 2004, pp. 288-292.

[10] Erdmann, M., Maedche, A., Schnurr, H.P., and Staab, S., "From Manual to Semi-Automatic Semantic Annotation: About Ontology-Based Text Annotation Tools," Linköping Electronic Articles in Computer and Information Science, Vol. 6, No. 002, 2001.

[11] Feigenbaum, E.A., "The Art of Artificial Intelligence: Themes and Case Studies of Knowledge Engineering," Proceedings of the 5th International Joint Conference on Artificial Intelligence, 1997, pp. 1014-1029.

國際替代計量

以本體為註記基礎應用於圖像資源管理

全文下載

主題瀏覽