3-dimensional error diffusion method based on edge detection for digital display devices

Abstract

The conventional error diffusion (CED) method adopted in digital display devices is simply modified from Floyd-Steinberg error diffusion method, which will degrade image quality by introducing nonlinear distortion (worm artifacts), linear distortion (sharpening), noise and the loss of image details. To overcome the shortcomings of the CED method, this paper proposes 3-dimensional (3D) error diffusion method based on edge detection. The proposed method diffuses quantization errors to the neighbor pixels of current frame and the neighbor pixel of the next frame. And the weights of error filters are dynamically adjusted based on the results of edge detection in each pixel's processing, which makes the weights coincide with the local edge feathers of input image. Simulation and experimental results show that the 3D error diffusion method can reduce worm artifacts and improve reproduction precision of image details.