摘要
文档图像作为图像的一种,在生活中的应用越来越广泛,然而其又不同于常规的文本文档或图像,它主要由具有特定含义的不同形状的字符串组成,其局部像素变化比较剧烈,高频分量相对丰富,采用常规的压缩方式很难获得较高的压缩率。常用的压缩方式没有考虑文档图像的特殊性,因而压缩性能有限。本文针对文档图像的特点,采用分块匹配的方法对文档图像进行压缩,即按照特定的规则对整幅图像进行分割,然后将分割的块图像进行分类并编码,从而在二维空间上消除了文档图像的相关性,获得了远高于常规无损压缩方式的压缩率。文中对分块匹配算法进行了描述,并对其性能进行了理论分析和仿真。
Document Image is a kind of image which is used widely in our life,while it is different from the general images and texts,its local pixels change quickly and it has abundant high frequency,so that it is difficult to get a higher compression rate with the general compression methods.Using general compression methods to compress document images has poor performance because the particularity of the document image is not considered.According to the characteristics of Document Image,the method of Block Matching is used to get a much higher compression rate than the general compression methods which eliminate the redundancy via two dimensions.We first split a document image with a specific method and then classify and encode the image blocks.We describe the method of Block Matching in detail and analyze the performance,and then we show the result of simulation.
出处
《计算机工程与科学》
CSCD
北大核心
2011年第7期45-49,共5页
Computer Engineering & Science
关键词
文档图像
分块匹配
无损压缩
document image
block matching
lossless compression
