摘要
当DNA图谱间的互信息量发生变化时,采用传统的算法进行图谱中的破损区域识别时,DNA图谱进行自适应分割不准确,提取的破损区域不完整,导致破损区域识别精确度低、效果差的问题。提出改进大津算法的DNA图谱中的破损区域识别方法。先利用大津法得到DNA图谱的一个整体阈值,并做为DNA图谱的初始分割阈值,以DNA图谱的整体互信息量最大为基础来计算分割图谱与初始图谱间的互信息量,并在互信息量达到最优值时对DNA图谱进行自适应分割,利用分割图谱与初始图谱间对比度、相关性和局部平稳特征参数,提取DNA图谱中的破损区域,进而有效的完成了DNA图谱中的破损区域识别。仿真结果证明,改进的识别方法对DNA图谱中的破损区域识别精确度高,分割效果好。
An identification method for damaged area in DNA map is presented based on improved Dajin algorithm. Dajin method is used to get a whole threshold of the DNA map, which is treated as the initial segmentation threshold. The mutual information between the partition map and the initial map is calculated by maximizing the overall mutual information of the DNA map, and the DNA map is adaptively segmented when the mutual information achieves the optimal value. The contrast ratio, correlation and local smooth characteristic parameters between the segmentation map and the initial map are used to extract the damaged area of DNA map, thus the identification of DNA damaged area is effectively completed. SimuLation proves that the proposed identification method has high precision and good segmentation effect.
出处
《计算机仿真》
CSCD
北大核心
2016年第4期360-363,共4页
Computer Simulation
关键词
大津法
阀值
互信息量
Dajin method
Threshold
Mutual information
