基于K-means聚类的数字半色调算法

Digital halftoning algorithm based on K-means clustering

数字半色调是在二值设备或多色二值设备上实现图像再现的一门技术,提出将K-means聚类法应用在数字半色调技术中。算法中应用人类视觉系统模型(HVS)和印刷模型最大限度减少原始灰度连续调图像和半色调图像之间的视觉误差;利用K-means聚类法将灰度图像划分成聚类分区,在每个聚类分区应用最小平方法(least-squares)最小化二值半色调图像和原始灰度级图像之间的平方误差,所构造的半色调算法与基于模型的最小平方法(LSMB)算法相比,随着聚类分区的增加,图像平滑且边缘清晰度增加,尤其是在图像细节部位。与LSMB算法比较,该算法的均方误差值有所降低,而权重信噪比和峰值信噪比提高了0.2～2 dB,模拟实验结果验证了算法的有效性。

Halftoning is a method for creating the illusion of continuous tone output with a binary device.This paper applied K-means clustering method to digital halftoning.The algorithm applied both a printer model and a model for the human visual system（HVS） to furthest minimize the perceived error between the continuous original image and the halftone image.Firstly,the method partitioned the gray image into two,three and four regions using K-means image segmentation method.Each clustering region used the least-squares model-based（LSMB） algorithm.Analysis and simulation results show that the proposed algorithm can produce better image smoothing and edge sharpness,especially the parts of image detail while increasing the number of clustering.Compared with the LSMB algorithm,it decreases the mean square error（MSEv） performance for the proposed algorithm,increases the weighted signal-to-noise ratio（WSNR） and the peak signal noise ratio（PSNR） performance for proposed algorithm by 0.2 to 2 dB.Experimental results indicate the effectiveness of the proposed algorithm.