基于图像特征信息的数字半色调方法

Digital halftoning method based on characteristic information of image

绝大部分数字半色调方法是通过图像的局部再现实现图像的整体再现,该技术存在细节信息少,特征不明显等不足.为此,提出基于图像特征信息的数字半色调方法.首先,选取合适的图像特征函数,构造绘图图元金字塔滤波器(其像素灰度值大于等于255)和权值金字塔滤波器;其次,利用绘图图元滤波器计算绘图图元的数量;最后,利用权值滤波器从上到下逐层排列绘图图元,直到该滤波器的最低层,即可得到再现图像.实验结果表明,根据用户的不同需求设置参数,能够再现图像不同程度的特征信息;再现图像不仅规律性纹理很少,而且能够充分再现图像的细节信息.

In most existing methods, some useful characteristic information and detail information were deficient render, because whole rendering and edge rendering of image were realized by local rendering. A new digital halftoning method based on characteristic information of image was developed. First, the image eigenfunction are originated, drawing primitives bandpass pyramid filter and weighted-bandpass pyramid filter are constructed. Second, the quantities of drawing primitives are calculated by the top element of drawing primitives bandpass pyramid filter. Finally, drawing primitives are arranged step by step from the top to the lowest of bandpass pyramid filter, as well as, the resultant image is gained. Lots of comparative experiments have improved that the rendering level of correlative characteristic of image is enormously raised according to user different requirement and different parameters, that the resultant digital image with new method is free from regular patterns, that some useful characteristic information and detail information are fully rendering.