Abstract
Most of the conventional image cognition methodologies are based on the geometrical singular points extraction of an image. This method has problem that requires the definition what is the singular point of image also the way how to extract the singular point.
To remove this difficulty, we have proposed the Eigen pattern method that was based on color information of images. Even though Eigen pattern method has been successfully applied the image cognition, image diagnosis and so on, it has been pointed out that the Eigen pattern method lost the original geometrical information of target image, i.e., geometrical complexity. To overcome this problem, this paper proposes that the geometrical information or complexity of image is represented by means of the modified Fourier power spectrum histogram.
