摘要
分别用灰度级直方图统计距和8邻域像素灰度差值法两种方法讨论唐卡图像纹理复杂性特征。用大量唐卡图像和一般图像进行对比实验,实验表明两类纹理量化方法均证明唐卡图像比一般图像纹理复杂度更高,尤其是8邻域灰度差值法充分考虑到像素间相对位置信息,用来刻画图像纹理复杂度更具优势。
Discusses the texture complexity of thangka images by using two methods: statistical distance based on histogram of grayscale and gray value difference method of 8 neighborhood pixels. Compares a large number of thangka images and general images for experiments, which shows that the texture complexity of thangka images is higher than that of general images by two kinds of texture quantization methods. In particular, the gray value difference method of 8 neighborhood pixels takes full account of local relative position between pixels, so it is more advantageous to describe the complexity of image texture.
作者
张春玉
ZHANG Chun-yu(School of Information Engineering,Xizang Minzu University,Xianyang 712082;Xizang Key Laboratory of Optical Information Processing and Visualization Technology,Xianyang 712082)
出处
《现代计算机》
2018年第21期25-28,共4页
Modern Computer
基金
西藏自治区自然科学基金重点项目(No.2016ZR-MZ-01)
关键词
唐卡图像
纹理复杂性
统计距
8邻域
差值法
Thangka Image
Texture Complexity
Statistical Distance
8 Neighborhood
Difference Method