摘要
维文联机手写单词中有大量的延迟笔划,在进行特征提取之前需要对延迟笔划进行探测。先采用精确基线查找的方法找到基线的位置,再依据基线信息划分成一定的区域,最后根据区域信息和维文延迟笔划的特点,采用多信息融合的方式去探测维文延迟笔划。对收集到的1088个维文手写体样本进行探测实验,正确探测率达到97.3%。实验结果表明,该方法是有效的,能够很好地探测出维文的延迟笔划。
There are many delay strokes in Uyghur online handwriting,the delay strokes should be detected before the feature extraction.In this paper,we firstly adopt accurate baseline search method to find the position of baseline,then divide the stroke into certain areas according to baseline information,and lastly detect the delay strokes of Uyghur using multi-information integration approach according to area information and the characteristics of Uyghur delay strokes.Detecting test has been carried out on 1088 samples of Uyghur handwriting collected,the correct ratio of detection achieves 97.3%.Experimental results show that this method is effective,and can better detect the delay strokes of Uyghur.
出处
《计算机应用与软件》
CSCD
北大核心
2012年第11期187-189,共3页
Computer Applications and Software
基金
新疆维吾尔自治区自然科学基金项目(200821124)
