An Improved Error Diffusion Algorithm

Halftoning is the rendition of continuous-tone pictures on binary devices, which iswidely used in printing, displaying and press fields. It bases on low pass filteringcharacteristics of human eyes and image characteristics and uses computer tools andmathematical tools to realize the reconstruction of images. With people’s higherdemands for the quality of images and the performance of digital hard-copy devices,how to improve halftone images has become one of the key research subjects.To solve the drawbacks in the standard error diffusion algorithms, such asdirectional hysteresis, fuzzy edges and poor contrast existing, this paper proposes animproved halftoning algorithm based on an improvement on the texture and feedbackof human visual errors. First, apply Floyd-Steinberg error diffusion to obtain the initialhalftone image. Second, scan the pixel of initial halftone image one by one and split itinto some blocks, which is as small as that of a bad texture, and then compare themand negate them every other line if they are bad textures. After that, calculate the visualerror with HVS model and new continuous-tone images by adding the feedback ofvisual error to original images are obtained. Finally, error diffusion procedure isexecuted again to get the final halftone images. Our method is utilized to test varioustypes of images. Simulation results show that our method is superior to the existingdigital halftoning methods on the aspect of subjective sense and objective evaluation.