基于骨架平滑与均匀膨胀的女书轮廓字形生成方法  被引量:2

Generation Method of NüShu Counter Font Based on Skeleton Smoothness and Uniform Dilation

在线阅读下载全文

作  者:徐凌[1] 王江晴[1] 李波[1] 

机构地区:[1]中南民族大学计算机科学学院,武汉430074

出  处:《科学技术与工程》2014年第34期229-234,共6页Science Technology and Engineering

基  金:国家自然科学基金面上项目(60975021);湖北省自然科学基金一般项目(2012FFC13401)资助

摘  要:针对手写女书字符目前尚未规范化,提取轮廓不平滑的问题,提出了一种笔画粗细均匀的女书轮廓字形生成方法,包括骨架提取、毛刺去除、骨架膨胀和轮廓提取等四个步骤。针对女书字符多呈菱形,弧笔和斜笔居多的特点,提出了一种单像素骨架提取算法;该算法结合八邻域模板邻域像素值分析与索引表遍历查找的思想,获得的骨架可以保持平滑性与骨架整体的单像素性。针对少量骨架结果伴有部分毛刺的情况,采用Rutoviz相交数分析法和Freeman方向链码法,设定不同的长度阈值和级别进行分级处理,有效去除毛刺。然后采用形态学运算对骨架进行均匀膨胀,得到笔画粗细均匀的女书字符,并采用边缘梯度和边缘跟踪的方法提取字符轮廓和生成轮廓字形。实验结果表明,该方法能快速、有效生成笔画粗细均匀、轮廓平滑的女书轮廓字形,对其他手写字符具有一定适用性。For the lack of standardization of handwritten NüShu characters and the un-smoothness of contour extracted,a method to generate the NüShu contour font of uniform stroke weight was proposed. The method included four steps of skeleton extraction,burr elimination,skeleton dilation and contour extraction. There were the characteristics of rhombus shape and arc or oblique strokes in the NüShu characters,so a single-pixel skeleton extraction algorithm was proposed. The pixel analysis of eight-neighbor template was combined with the traversal search of index table. On account of this algorithm the skeleton possessed the characteristics of smoothness and overall singlepixel. For small amount of skeletons with a few burrs,Rutoviz crossing-number and Freeman direction chain code were used to eliminate burrs in different level and length thresholds effectively. Morphological operation was used to dilate the skeleton uniformly and get the NüShu characters of uniform stroke weight. Edge gradient and edge tracking were used to extract the character contour and generate contour font. The experimental results show that this method can generate the NüShu contour font of uniform stroke weight and smooth contour fast and effectively. It has certain applicability for other handwritten characters.

关 键 词:女书 骨架提取 单像素 毛刺去除 骨架膨胀 轮廓字形 

分 类 号:TP391.1[自动化与计算机技术—计算机应用技术]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象