论文部分内容阅读
针对脱机手写维吾尔文本行图像中单词切分问题,本文提出了FCM融合K-means的聚类算法。通过该算法得到单词内距离和单词间距离两种分类。以聚类结果为依据,对文字区域进行合并,得到切分点,再对切分点内的文字进行连通域标注,进行着色处理。本文以50副不同的人书写的维吾尔脱机手写文本图像为实验对象,共有536行和4002个单词,正确切分率达到80.68%。实验结果表明,该方法解决了手写维吾尔文在切分过程中,单词间距离不规律带来的切分困难的问题和一些单词间重叠的问题。同时实现了大篇幅手写文本图像的整体处理。
Aiming at the problem of word segmentation in off-line handwritten Uyghur text lines, this paper proposes a clustering algorithm based on FCM fusion K-means. Through the algorithm, the word distance and word distance are two kinds of classification. Based on the clustering result, the text area is merged to obtain the segmentation point, and then the text in the segmentation point is marked with the connected domain, and the coloring process is performed. In this paper, Uyghur off-line handwritten text images written by 50 different people were used as experimental subjects, with a total of 536 lines and 4002 words, with a correct segmentation rate of 80.68%. Experimental results show that this method solves the problem of handwriting Uighur segmentation in the segmentation process, the problem of segmentation difficulties caused by the irregular distance between words and the overlap between some words. At the same time to achieve large-scale handwritten text image of the overall treatment.