The Study Of Error Diffusion Digital Halftoning Technique

Digital halftone technology，which based on the visual characteristics of the human eye andthe fineness of the image features, is an image reproduction technology in binary devices orlimited grayscale devices with tools such as mathematics, computer. The technology is widelyused in printing, textile, digital image compression and storage, image transmission, medicineand other fields. Studing digital halftone technology has important practical significance andapplication value.Currently the error diffusion halftoning is one of most popular digital halftoning. The paperdetailedly analyses of the significance and the development status of the halftone technique, andit is necessary to research in this area. The whole research begins with researching in the digitalrepresentation of the image, the visual characteristics of the human eye, the basic principles ofdigital halftone technology and briefly classification of digital halftone technology, such aspoint-processing algorithms, the field processing algorithms and iterative processing algorithms.In addition, the paper gives quality evaluation criteria of digital halftone image.At the same time, the paper detailedly analyses of several classic error diffusion method,such as the F-S error diffusion algorithm, point diffusion algorithm, model-based error diffusionmethod, blue noise error diffusion method and edge enhanced error diffusion method. And thenthe paper compares the proliferation of this classical error diffusion algorithm with each other.Subsequently, the paper thoroughly analyses the performance of error diffusion algorithms,related to the principle of the error diffusion, how to selecte the pixel scan path, how to select thequantization threshold, how to set the coefficients of the error filter, the deformation as well asstability of the system.On the basis of the foregoing study, the paper gives an improved algorithm, which is thedual feedback error diffusion algorithm based on edge detection. And then the paper explainsthe principle of the improved algorithm in detail. Meanwhile, how to extract edge pixels, how toset the coefficients of the error diffusion filter and the feedback filter, how to select thequantization threshold, and how to realize the whole improved algorithm have been made adetailed analysis. Finally, the paper compares the similation results of the improved algorithm tothe similation results of the F-S algorithm. Besides, the paper offers their objective evalution paraments to readers. It is easy to see that the edge of halftoned image has been enhanced and themoire phenomenon has been reduced. The image acquired from improved algotithm is clearerthan the one from F-S algorithm.