國立臺灣大學電機工程學系暨研究所陳銘憲2006-08-232018-07-062006-08-232018-07-062003-11-30http://ntur.lib.ntu.edu.tw//handle/246246/29480現今由於多媒體技術的長足進步,互 動式多媒體資訊檢索系統廣泛地發展與 應用在日常生活之中。這些系統不但要能 有效率的管理多媒體物件,亦必須擁有自 動辨識物件的能力。尤以視訊媒體的蓬勃 發展,視訊處理與利用的技術亦趨重要。 視訊自動分類、視訊摘要、語意檢索等技 術將成為視訊勘測上不可或缺的利器。 我們認為視訊勘測在未來之發展深 具潛力,其重要性亦與日俱增。因此在此 計畫中,我們致力於以模型物件為基礎之 互動式視訊勘測技術之研究設計。我們欲 涵蓋之範圍包括:(1) 利用相關視訊處理 技術,設計以模型物件為基礎之互動式視 訊勘測系統;(2) 針對視訊資料的特性, 研究並設計視訊物件辨識與分析所應具 備之架構;(3) 針對視訊勘測 (video mining) 的需求,研發三種關鍵性視訊勘 測模組: 自動分類 (auto classification) 、視訊摘要 (video summary) 、語意檢索 (semantic retrieval);(4) 依據三種勘測技術所獲 得之知識,探討如何將其應用於日益普及 的互動式多媒體資訊檢索系統,以提供更 便利的個人化服務。At present, based on the tremendous progress of multimedia technology, interactive multimedia information systems of the real world are being widely developed and applied. The systems are required not only to effectively manage multimedia objects, but also to present the ways of automatic recognition of objects. Particularly, with the fast growth of video sources, efficient video processing and utilizing techniques become more and more important. Automatic video classification, video summary, and semantic retrieval methods provide powerful support to meet this demand of video data mining. With its potential markets, we believe that the video data mining will be more and more important in the future. In the project, we explore the design of an interactive model-based video mining and auto classification system. Our working items include (1) Based on the advanced video processing technologies, we study an interactive model-based video mining system; (2) Based on the characteristics of video data, we explore the framework of video objects recognition and analysis; (3) Based on the demand of video mining, we study three video data mining capabilities, which include automatic classification, video summary, and semantic retrieval; (4) Based on the knowledge obtained in the video mining system, we aim to provide every customer the personalization services of multimedia information from anywhere at anytime.application/pdf36804 bytesapplication/pdfzh-TW國立臺灣大學電機工程學系暨研究所以模型物件為基礎之互動式視訊勘測技術Model-based Video Mining and Auto Classificationreporthttp://ntur.lib.ntu.edu.tw/bitstream/246246/29480/1/912622E002044CC3.pdf