1 September 1990 Image coding by edge primitives
Author Affiliations +
Proceedings Volume 1360, Visual Communications and Image Processing '90: Fifth in a Series; (1990) https://doi.org/10.1117/12.24174
Event: Visual Communications and Image Processing '90, 1990, Lausanne, Switzerland
The visual system is most sensitive to structure. Moreover, many other attributes of the scene are preserved in the retinal response to edges. In this paper we present a new coding process that is based on models suggested for retinal processing in human vision. This process extracts and codes only features that are preserved by the response of these filters to an edge (edge primitives). The decoded image, obtained by recovering the intensity levels between the outlined boundaries, using only the edge primitives, attains high structural fidelity. We demonstrate that a wide variety of images can be represented by their edge primitives with high compression ratios, typically two to three orders of magnitude, depending on the target. This method is particularly advantageous when high structural fidelity representation must be combined with high data compression.
© (1990) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Rachel Alter-Gartenberg, Rachel Alter-Gartenberg, Ramkumar Narayanswamy, Ramkumar Narayanswamy, } "Image coding by edge primitives", Proc. SPIE 1360, Visual Communications and Image Processing '90: Fifth in a Series, (1 September 1990); doi: 10.1117/12.24174; https://doi.org/10.1117/12.24174


Back to Top