WANG Qi,CUI Rongyi*.Layout global complexity evaluation method based on byte stream information entropy[J].Journal of Yanbian University,2019,45(02):136-140.
基于字节流信息熵的版面全局复杂度的评估方法
- Title:
- Layout global complexity evaluation method based on byte stream information entropy
- 文章编号:
- 1004-4353(2019)02-0136-05
- 分类号:
- TP391
- 文献标志码:
- A
- 摘要:
- 以图文要素构成的word 2003版面存储文档为研究对象,提出了一种利用信息熵评估版面文档复杂度的方法.首先,从图像和文本存储特点出发,提出一种利用文件字节流信息熵度量版面全局复杂度的方案; 其次,将文件视为信源,每个字节视为信源符号,以二进制方式读取文件,然后根据字节相关性,采用N 次扩展信源计算信息熵; 最后,通过实验验证表明,本文方法切实可行,给出的版面全局复杂度定量描述不仅能很好地符合人的视觉直观感受,而且能够为版面数据可压缩性提供依据.
- Abstract:
- The word 2003 layout document composed of graphic elements is taken as the research object, and a method for evaluating the complexity of layout documents by using information entropy is proposed. Firstly, based on the characteristics of image and text storage, a scheme for measuring the global complexity of layout using file byte stream information entropy is proposed. Secondly, the file is regarded as the source, and each byte is regarded as the source symbol. Read files in binary mode, and then the information entropy is calculated by N times based on the byte correlation. Finally, the experimental results show that the method is effective and the quantitative description of the global complexity of the layout is not only well matched human visual perception, but also can provide a basis for layout data compressibility.
参考文献/References:
[1] GUO X, ASANO C M, ASANO A, et al. Visual complexity perception and texture image characteristics[C]//International Conference on Biometrics and Kansei Engineering. Washington, D C: IEEE Computer Society, 2011:260-265.
[2] 王浩.基于颜色和纹理特征的图像复杂度研究[D].长春:长春理工大学,2016.
[3] 高振宇,杨晓梅,龚剑明,等.图像复杂度描述方法研究[J].中国图像图形学报,2010,15(1):129-135.
[4] 党兴.复杂的中文文档图像版面分析研究[D].苏州:苏州大学,2010.
[5] PETERS II A, STRICKLAND R. Image complexity metric for automatic Target recognizers[C]//Automatic Target Recognizer System and Technology Conference, October. Silver Spring: Naval Surface Warfare Center, 1990:1-17.
[6] ZOU J, LIU C C. Texture classification by matching co -occurrence matrices on statistical manifolds[C]//10th IEEE International Conference on Computer and Information Technology(CIT 2010). Washington, D C: IEEE Computer Society, 2010:1-7.
[7] 傅祖芸.信息论[M].北京:电子工业出版社,2007:190-219.图1 数据处理示意图
备注/Memo
收稿日期: 2019-03-26 *通信作者: 崔荣一(1962—),男,博士,教授,研究方向为模式识别、智能计算.
*基金项目: 吉林省自然科学基金资助项目(20140101186JC); 国家语委科研立项基金资助项目(YB135-76)