Estimating dither thresholds from the average of halftone dots

Conventional approach to dither threshold estimation is based on the full dot patterns (FDP´s), which are selected from the given halftone. The main drawback of the approach is some FDP´s may be missing from the halftone. We propose to estimate the threshold from the average of all the halftone dots. Our method can additionally overcome the missing FDP´s arising from the nearly-constant gray values present in the image but not gathered enough to cover a “full dot”. Experimental results show that compared with the FDP-based approach, our method is better and less degraded by the increasing of threshold size.