欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
离线手写体的笔顺提取中用到的相邻笔段间最小曲率的假设,对于中文书法文字的笔顺提取并不完全适用,为此提出基于书法统计结构混合模型,通过该模型获得笔顺的新方法,其步骤是：（1）按照书法文字模型提取样本的相关特征,（2）按照输入样本使用粗糙选择获得候选样本集,（3）使用条件密度传播算法从已知笔顺的候选样本集中搜索出和输入样本最佳匹配的样本.对100个书法文字中的15个常用部首进行笔顺提取,基于最小曲率的Kalman滤波和新方法的准确率分别为42%和91%.
英文摘要：
The assumption of minimum curvature between adjacent line segments which is used to recover strokes order from off-line handwritings is not fully applicable for strokes order recovery from Chinese calligraphic handwriting. A new method based on calligraphic hybrid statistical-structural model to recover strokes order was proposed. The stages are：（1） the features of input sample are extracted according to Chinese calligraphy model （CCM）; （2） coarse selection is applied to find the candidate samples set; （3） CONDENSATION algorithm-conditional density propagation over time is performed to select the best matching sample from candidate samples set with known strokes order. By recovering strokes order from 15 commonly used radicals of 100 Chinese calligraphic handwritings,the accuracy of Kalman filter based on minimum curvature and that of the new method were 42% and 91% respectively.
同期刊论文项目
同项目期刊论文
Personalized retrieval of sports video based on multi-modal analysis and user preference acquisition
期刊信息
位置：