Optimization of Multilevel Error Diffusion Based on Pixel Distribution

To improve the visual effect of multilevel halftone image, a multilevel quantization interval division method based on pixel distribution and a dynamic error limits method are presented. First, the impact of quantization intervals on the mean square error (MSE) of the image is analyzed. According to that, quantization intervals optimization is presented to make the number of pixels in each interval equal based on the pixel distribution of original image. That optimization leads to uneven quantization intervals. Then, a dynamic error limit of error diffusion method is utilized to minimize the statistical error between the resulted image and the original one. The results show that the peak signal-to-noise ratio (PSNR) of proposed method is increased about 2 dB compared with the Floyd-Steinberg algorithm. And the efficiency of our quantization is nearly twice of Octree quantization algorithm. The resulted images receive acceptable visual effect.