Stability analysis of error diffusion

  • Authors:

  • Zhigang Fan

  • Affiliations:

  • Webster Research Center, Xerox Corporation, Webster, NY

  • Venue:
  • ICASSP'93 Proceedings of the 1993 IEEE international conference on Acoustics, speech, and signal processing: image and multidimensional signal processing - Volume V
  • Year:
  • 1993

Quantified Score

Hi-index 0.00

Visualization

Abstract

Error diffusion is an important technique for digital halftoning. In this paper the stability conditions for error diffusion is analyzed. A BIBS (Bounded Input Bounded State) type definition is adopted. The error difhsion systems can be classified into three types according to the filters associated with them. The Type I systems are stable. Provided the input values are bounded by the possible output states, the Type III systems are either stable or can be considered practically stable. The Type I1 systems are unstable generally. Some of them can be stable if the input ranges are small enough. However, their chaotic behavior might be unacceptable for many practical applications.