Research On Template Optimized Of Inverse Halftoning Technology For Look-up Table

Image halftone technology and the inverse halftoning technology is a branch of image processing technology.The Halftone process is implanted the noise into continuous image,and becomed it into binary image.Look-up table method is one of many inverse halftoning method art machine learning methods,calculation simple and can be processed in parallel advantage of look-up table method,and thus become one of the main technologies currently inverse halftoning methods.Aiming to find the look-up table optimal template selection,we depth study,and the main work is as follows:1)We proposed an algorithm to find the best Look-up Table template selection method based on orthogonal differential evolution algorithm.For optimal lookup table template selection,we use orthogonal differential evolution algorithm to find the optimal template.The establishment of a best template selection model with constraints,according to the model designed orthogonal experimental program,coding,crossover operator,mutation operator and constraints of individual operators,individual evaluation function.To improve the population differential evolution algorithm optimized for speed so as to enhance the convergence rate results in differential evolution algorithm using orthogonal cross-step method.Finally,experimental results show that the algorithm is compared with genetic algorithm and simulated annealing algorithm has global convergence speed,stable result advantages.2)A human visual system based on optimal selection model lookup table template structural similarity is presented.For existing lookup table template selection algorithm in judging the degree of bad template without considering human visual characteristics,Thus the design template selection model which based on human visual system.In this model,the use of 5/3 wavelet based on lifting structure will break down the image into four regions,the image information for the four regions retained,classic structural similarity and similarity to judge based on graded structure,synthesized the results of the final evaluation will be forthe final result.Experiments show that the model obtained template when rebuilding the halftone image more in line with the visual characteristics of the human eye.