文章基本信息

标题：Digital Contour Segmentation Based on Statistical Straightness
本地全文：下载
作者：Liang Jia ; Jiuzhen Liang
期刊名称：International Journal of Signal Processing, Image Processing and Pattern Recognition
印刷版ISSN：2005-4254
出版年度：2016
卷号：9
期号：8
页码：295-310
出版社：SERSC
摘要：A novel algorithm for segmenting digital contours in edge images to digital straight segments (DSS) is proposed in this paper. While exploring contours, the algorithm maintains a dynamically-updated histogram of directions between every two adjacentcontour pixels which are considered connected with respect to a generalized concept of connectedness. Directions of pixels are compared with the direction of the greatest value in histogram, and the ones of differences less than a given threshold are collected as the candidates of DSS. Finally, DSS is split and merged based on the analyses of the properties associated with the histogram such as balance and modality. Implementation of the proposed algorithm is visually introduced by using UML activity diagrams, and the experimental results of the artificial and real-life images are illustrated and discussed. Analyses of the experiments showed the validity of the segmentations and the bottleneck of the proposed algorithm.
关键词：Digital straight segment; Line segment; Feature extraction; Digital image ;processing