测度熵在图像纹理分析中的应用  

AN APPLICATION OF MEASURE ENTROPY IN IMAGE TEXTURE ANALYSES

作　　者：孟勇[1] 洪丹辉[1] 毛丹[1] 

机构地区：[1]复旦大学教学所,上海200437

出　　处：《计算机应用与软件》2000年第8期52-56,共5页Computer Applications and Software

摘　　要：本文的主要目的是把测度熵应用在图像的纹理分析中。测度熵比基于统计的方法要可靠、方便。把它应用于对纯卡通图、扫描卡通图、混合图以及照片的识别,大大提高了图像自动生成绣花针脚软件的价值。另外,测度熵还能有效地检测出图像纹理的光滑度,对绣花软件来说,我们就可以根据光滑度从针脚库中自动选出合适的针脚,使得绣图更加逼真、自然。The purpose of this paper is to apply measure entropy in image texture analyses. It's important to find an efficient method for image texture analyses in some field of image processing. We realize that measure entropy is a very useful tool and applied to an embroidery software which can auto-generate stitch,and make this product really realize that it can auto-classify pure cartoon picture,scanned cartoon picture, mixed picture and photo. Thus it is easy to choose different ways to process respectively. Additionally, measure entropy also can detect the smooth degree of image texture efficiently. To embroidery software, we can choose suitable stitch library according to the smooth degree and make embroidery picture more vivid and natural. Of course, measure entropy is very useful in other fields.

关 键 词：测度熵 图像纹理分析 图像处理 计算机 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]