期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2017
卷号：95
期号：9
出版社：Journal of Theoretical and Applied
摘要：Indian Cultural dance poses are segmented using two algorithms and their performance is estimated quantitatively. A data set of various dance poses are captured under controlled environments. The image dataset consists of 100 dance mudras of kuchipudi dance poses. Marker controlled Watershed and super pixel based simple linear Iterative clustering (SLIC) algorithms are proposed to accomplish the segmentation task. The results of both the algorithms show that SLIC based super pixel segmentation outperforms the watershed algorithm. Further, the results of the segmentation can be used for classification of various Indian dance forms.
关键词：Dance Image Segmentation; Watershed Algorithm; Super Pixels; Simple Linear Iterative Clustering.
