摘要
针对现有维吾尔文笔迹特征提取方法缺乏旋转不变性导致识别存在偏转的样本效果较差,以及尺度不变特征变换(SIFT)方法用于维吾尔文笔迹鉴别存在不足的问题,提出一种基于特征融合、具有旋转不变性的鉴别方法。该方法首先提取笔迹图像的SIFT特征,再计算局部窗口特征,并将两者融合对旋转角度不同的笔迹样本进行鉴别。实验证明,该方法能有效克服笔迹样本旋转对识别率造成的影响,是一种简单、实用、识别率较高的维吾尔文笔迹鉴别方法。
Due to the lack of rotation invariance characteristics of existing methods for Uyghur handwriting verification which lead to poor results for rotated handwriting samples,and shortage of using scale invariant feature transform(SIFT),this paper presented a method based on features combination with rotation invariance characteristics.The proposed method firstly extracted SIFT feature,then calculated local window feature and combined both of them before being used for handwriting samples of different rotation angles.Experiments show that the proposed method can effectively overcome the impact on identification of the rotation of handwriting samples and is a simple,practical method with favorable performance for Uyghur handwritings.
出处
《计算机应用研究》
CSCD
北大核心
2012年第11期4342-4344,共3页
Application Research of Computers
基金
国家自然科学基金资助项目(61173130)
关键词
笔迹鉴别
维吾尔文笔迹
尺度不变特征变换
局部窗口特征
旋转不变性
writer identification
Uyghur handwritings
invariant feature transform
local window feature
rotation invariance