近年來在科技快速的發展及網際網路的普及化下,學校一改以往以言語描述為主的教學方式,轉變為多樣化的影音數位教學方式。然而以往一般實物教學通常是與口頭講授配合,很少能夠與多媒體教材緊密結合。因此如有一套實物教學的輔助平台,在課堂上進行實物教學時可以用來協助實物教材與多媒體教材之結合,且可讓學生在課後之餘利用網路對該平台進行檢索學習,不但能進一步提昇實物教學的教學成效,也能提昇學生課後學習的動機。本研究希望建置一套通用的實物教學輔助平台,該平台可讓教師於課堂上應用,且可供學生在課後進行教材影像查詢及學習。在教材輸入方面,本研究所提出之系統提供Java Application供教師自行建立教材特徵資料庫,且系統會自行根據教師所提供之教材特徵資料庫的樣本特性,自動的選取適用於該教材特徵資料庫之特徵值以及比對演算法,讓對影像處理技術較不熟悉之教師也可簡便的使用該平台,且學生也可透過Java Application及Web-based兩種方式應用該實物教學輔助平台。期望本系統能讓實物教學之成效以及學生之學習意願再次提升。
Because of the rapid progress of computer techniques, images and videos and other multimedia information are incorporated into the object teaching. The object teaching system can benefit not only the in-class instruction but also after-school self-learning. Commonly, the database of the system is setup by the developers. It is difficult for the users to increase or setup the contents of the database. In this thesis, we propose a user-friendly object teaching system. Users can setup the database by providing the samples or images of the samples. The system will automatically select the best features and the most suitable classifier. We develop the system in Java and implement the system in both web-based platform and independent application system. Testing with 12 kinds of common seashells, our system can reach 93% recognition rate without any tuning. We hope this system can effectively support class-teaching and promote the students’ will of self-learning.