This paper describes a contour-based representation of images. More particularly, we discuss a method of representing image data which is a resolution-independent scheme and can provide an interactive viewing capability for every user. In this method, a two dimensional (x,y) image is regarded as a curved surface in a three dimensional space having luminance/chrominance value of an image measured along z-axis. And an image is represented using contour lines which are contours of cross-sectional areas made by slicing a curved surface perpendicularly through z-axis. We also discuss a scheme for reconstructing images based on the proposed contour-based representation. Within the framework of image representation, contour lines are stored in a vector-coded format which is independent of the pixel resolution. Consequently a scalable representation algorithm can be realized, and it is possible to reconstruct image in quality which is in compliance with various kinds of user's requests.
"Contour-based image representation", Proc. SPIE 2915, Video Techniques and Software for Full-Service Networks, (17 January 1997); doi: 10.1117/12.263398; https://doi.org/10.1117/12.263398