圖像所包含的資訊非常的豐富,圖像內的各個物件間的特徵及空間關係或
者是物件的顏色.本論文的目的在研究如何利用這些圖像內的資訊來取得
一些圖像的特徵,以提供存取搜尋圖像資料庫之用.而本論文的研究內容分
為下列三個部份:首先是介紹如何取得圖像內特定物件的方法;然後就所取
得的物件內容,分析其顏色的分佈狀況,並且針對其顏色的RGB的分佈情形,
對照表示對於顏色的等感覺度,等差間距的顏色色相色立圖,以建立該影像
的特徵值;最後以此顏色特徵為索引值,將圖像建入圖像資料庫,並以此索
引值為日後存取與搜尋資料庫之用.

The information contained in an image is rich yet inexplicit. It
can be represented by the spatial relationship among the
objectsin the image. or by the characteristics of the image.
This thesisinvestigates methods for extracting characteristics,
in particual,color information,from image in order to search the
images fromimage database.This thesis proceeds in three steps:
First,introduce the methods for extracting objects from an image
. Second,analyze the RGB distribution of the objects of
interest,and match the distribution with the colordiagram to
contain a characteristic value for the objects as well asthe
image.Third,insert the image to the image database based on its
characteristic value.The classification within the color is
converted into an RGB table diagram.so that the diagram can be
easily matched withthe RGB distributions of images.