WLSMB Halftoning Based on Improved K-means Cluster Algorithm Using Direct Binary Search

Abstract

This work employs the well known weighted least squares method to optimization to produce halftone images using improved K-means clustering theory. Our algorithm applies to both a printer model and a model for the human visual system (HVS). In this algorithm, the improved K-means clustering method is used to segment an image several regions. In the halftone process, each clustering uses the weighted least-squares model-based(WLSMB) algorithm by use of direct binary search iterative method to obtain halftone image. Analysis and simulation results show that the proposed algorithm produces better gray-scale halftone image quality when we increase the number of clustering with a certain range and outperforms least-squares model-based algorithm in the PSNR (Peak Signal Noise Ratio), WSNR (Weighted Signal Noise Ratio) criteria.