Digital Halftoning is a gray level rendition in the device which has limited color representation capabilities. To overcome this limit, many researchers have proposed so many halftoning algorithms. Error diffusion dither algorithm is one of those which shows good halftoned image. However, it is not efficient for real-time halftoning because of its large computation amount. The blue noise mask and max-energy mask algorithms are alternative of this algorithm, but it has lower image quality and correlation artifacts.
This thesis proposes a new algorithm using optimally positioned halftoned pattern representing constant gray level image for human visual model. The contrast sensitivity is used as criterion of the optimal pattern generation. Also, the proposed algorithm uses the binary patterns for real image halftoning which combines the optimal constant pattern and natural image.
The proposed algorithm needs small amount of computation. It is fast algorithm for optimization of binary pattern and generation of real halftoned image. This is fit to hardware implementation very well.
