文章基本信息

标题：Performance Analysis of Comparison between Region Growing, Adaptive Threshold and Watershed Methods for Image Segmentation
本地全文：下载
作者：Erwin ; Saparudin ; Adam Nevriyanto 等
期刊名称：Lecture Notes in Engineering and Computer Science
印刷版ISSN：2078-0958
电子版ISSN：2078-0966
出版年度：2018
卷号：2233&2234
页码：157-163
出版社：Newswood and International Association of Engineers
摘要：Image Segmentation with region growing technique, clustering neighbor’s pixels and similar seed points otherwise adaptive thresholding create fixed blocks and find appropriate threshold values. Using images from Berkeley Segmentation Dataset (BSDS) is BSDS300, including 300 grayscale’s images and 300 color’s images, each of which has 200 training images and 100 testing images. The results are average performance measurement of precision, recall and FScore each of which has 0.437, 0.665, 0.525 for region growing method and 0.30, 0.525, 0.73 for adaptive thresholding method. While using watershed techniques, we obtained following values: 0.258, 0.488, and 0.333.
关键词：Segmentation; Region Growing; Adaptive; Thresholding; Watershed.