Scalable and high-precision function-approximated images and its secure coding

Function-approximated images are suitable for transmission over the networks, because they can be scaled on the sidelines of various resolutions of network terminals with relatively small amount of information. In this work, we use the Fluency function-approximation method that utilizes the Fluency information theory. The theory can deal with images in scalable and high-precision. Transmitted images may contain critical information that must be protected from malicious information tapping. The security, however, depends on the implementations of lower layers of the network protocol stack that are not available in many circumstances. In this paper, we propose a new image description format that applies security to function-approximation coding. It utilizes the structured properties of function-approximation format to enhance its security by hiding the coordinates of components of images. The proposed method is a secure mechanism for image coding itself; therefore, it enables secure transmission of function-approximated images without requiring any secure protocols in the lower layers. Furthermore, being independent from the lower layers, it enables flexible combinations of security mechanisms to enhance the security.