Oversampling and Coarse Quantization for Signals

CSIC Building (#406), Seminar Room 4122.
Directions: home.cscamm.umd.edu/directions

The Error Diffusion Halftoning Algorithm:
Some Recent Stability Results
and Applications Beyond Halftoning

Dr. Chai Wu

Thomas J. Watson Research Center at IBM

Abstract:   Halftoning is the art of producing a picture which appears to have many colors using only a handful of colors. Almost all digital and analog printers use some form of halftoning; just look at any picture in a newspaper or magazine under a magnifying glass. Error diffusion is a popular technique for high quality digital halftoning. The purpose of this talk is to illustrate the versatility of error diffusion with applications beyond halftoning and to show its connection to interesting mathematical questions. I will present some recent developments in the theory and applications of error diffusion and related algorithms. In particular, I will present several stability results of error diffusion and show applications in a variety of areas besides image halftoning including online scheduling, image watermarking, steganography and data hiding, and image enhancement in twisted-nematic mode liquid crystal displays.

[LECTURE SLIDES]