Characteristic Extraction of Dynamic Image by Fourier Cosinusoidal Transform

Abstract

This paper concerns with a methodology of geometrical characteristic extraction from visualized dynamic image. At first, we briefly describe the color Eigen pattern method, which has been proposed by us to cognize the static images realizing the fully automatic security and product inspection systems. Since this color Eigen pattern method is based on the color information of distinct target image, then it is possible to cognize the target having the same color information. On the other side, our color Eigen pattern method could not distinguish the different targets having the same color information.
To overcome this difficulty, we propose here a geometrical Eigen pattern method based on the Fourier Cosinusoidal transform. As a result, it is revealed that fairly good cognition may be promised by our geometrical Eigen pattern method at the initial test stage.