一种自适应误差分散逆半调算法  

An adaptive error diffusion inverse halftoning algorithm

作　　者：郑海红[1] 王义峰[1] 曾平[1] 

机构地区：[1]西安电子科技大学计算机学院,西安710071

出　　处：《哈尔滨工业大学学报》2009年第11期235-237,共3页Journal of Harbin Institute of Technology

基　　金：国防科研基金资助项目(51416050301DZ0140);陕西省自然科学研究计划资助项目(2001x06)

摘　　要：针对误差分散类半调图,提出一种基于Laplacian(L)金字塔的自适应逆半调算法.通过研究半调噪声与误差分散核(半调核)的关系,引入噪声粒度反映不同半调核的影响;采用分治策略,利用L金字塔变换提取含噪高频,采用融合技术将半调噪声转化为易于处理的脉冲噪声,使用维纳滤波器和中值滤波器去噪.实验证明,该算法能自适应选择阈值,具有保图像边缘清晰、时空复杂度较低、峰值信噪比高等特点.For error diffused halftone images, an adaptive inverse halftoning algorithm based on Laplacian pyramid is proposed. By analyzing the relation between halftone noise and halftone kernels, noise granularity is given. With the idea of dividing and treating respectively, Laplacian pyramid transformation is used to extract noisy high-frequency information, fusion method is employed to convert halftone noise into tractable impulse noise, and Wiener filter together with median filter is chosen as denoising processor. Experiments show that the proposed algorithm is capable of selecting threshold adaptively and has the advantage of low implementation cost, good edge-preserving capability and high peak signal-to-noise ratio.

关 键 词：逆半调 误差分散 Laplacian金字塔 图像融合 脉冲噪声 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]