Abstract
Transmission and reception of the Image/Video digitized by regular connection of a high-speed network and the spread of the high spec. PC has become possible anytime and anywhere. However, the technology of searching the Image/Video to mean efficiently is not enough. As this factor, it is exactly the polysemy that an Image/Video has. In Image/Video retrieval, making metadata that extracts exactly the interface that generates correctly the Image/Video that one imagined as a query, and the feature that an Image/Video has is the most important, and the accuracy of this matching act on reference accuracy greatly. This research describes the classification of the new amount expression of the features using the parameter of a photo subject's depth information recognized in experience common, and a camera position and an angle, and the Image/Video using them to many natural pictures that these subjects should be solved.
