藏文字形轮廓特征优化识别提取仿真研究

Tibetan Character Recognition in the Glyph Outline Feature Extraction Algorithms And Simulation

对藏文字形轮廓特征优化识别提取,可以加深对藏文字的了解,提高藏文的艺术价值。进行藏文字形轮廓特征识别提取时,应将藏文字体骨架的重要特征信息提取作为基础,但是传统方法用利用直线和二次Bezier曲线对字形轮廓进行识别提取,难以保留藏文字体骨架的重要特征信息,降低了藏文字形轮廓特征优化识别提取效果。提出一种改进小波变换的藏文字形轮廓特征优化识别提取方法。该方法先将小波边缘检测和笔段合成法相融合,对图像藏文边缘和骨架轮廓进行检测,定义藏文笔划的初边缘选取规则,利用模糊熵选取藏文结构图像不同灰度层次的阈值,增强图像中不同灰度层次的边缘信息,定义藏文图像新的隶属函数和模糊增强算子,经过平滑处理后完成藏文字形轮廓特征优化识别提取。仿真结果表明,所提方法可以有效地识别提取出藏文字形轮廓特征,为研究藏文文字的字形优化提供了重要的依据。

It can deepen the understanding of Tibetan alphabet and improve its artistic value to make optimization recognition and extraction to glyph outline feature of Tibetan. The recognition and extraction should take the important feature information of font skeleton as the basis. Traditional method recognizes and extracts the glyph outline using straight line and quadric Bezier curve, it cannot reserve the important feature information of Tibetan font skeleton and reduces the optimization recognition and extraction effect. An optimization recognition and extraction method for glyph outline feature of Tibetan is proposed based on the modified wavelet transform. Firstly, the wavelet edge detection is integrated with method of stroke segment synthesis to detect the Tibetan image edge and skeleton outline and define the edge selection rule of Tibetan stroke. Then the fuzzy entropy is used to select the threshold value of different gray level in Tibetan structure image and enhanced edge information of different gray level in image. The new membership function and fuzzy enhancement operator of Tibetan image are defined. Finally, the optimization recognition and extraction for glyph outline feature of Tibetan is completed after smoothing. The simulation results show that the method mentioned above can recognize and extract the glyph outline feature of Tibetan.