Skip to content

matlab project for Digital Image Processing, use SLIC super pixel segmentation and maximum flow graph cut to build an interactive fore/background segmentation app

Notifications You must be signed in to change notification settings

ValMystletainn/SLICInteractiveCut

Repository files navigation

文件结构
脚本:
dynamicSLIC.m 	-- SLIC 迭代步骤的动态显示脚本,算法本身同mySLIC
mySLIC.m		-- SLIC 算法的实现
mySLICMannul 	-- 可追加手动种子点的SLIC算法,即SLIC的局部交互改进 
postProcess	-- 增强SLIC结果连通性的后处理算法


pdf:

homework-2017010910-吴文绪 -- 作业报告

data
x = 1,2,3
goodresultx.jpg 		-- 使用很大超像素数(1k量级)做SLIC,再有耐心的划分前背景后得到的最好结果,作为基准使用
lesspointGoodResultx.jpg	-- 使用超像素数(100左右)做SLIC,辅之以合适的其他分割增强算法后得到的较好结果

程序使用说明

图文并茂的说明可见报告中   界面设计和使用说明 一节
打开待处理图片后,在左上方设置合理的SLIC参数,包括超像素数k,距离权重m,是否需要后处理,以及参与SLIC的像素特征层。
设定完成后,点击SLIC按钮执行SLIC,中间即会显示超像素分割结果。
可选的,如果我们对部分分割结果不满意,可以对这些部分进行局部交互,手动添加种子点,再次进行SLIC,以得到一个较好的分区结果。
然后点选前背景,点击lazysnapping按钮即可在右侧得到分割结果,同样,也可以根据结果,增删前背景点,以得到所期望的分割结果。

显示SLIC迭代过程,点击中间的show process按钮,会在一个新图窗中显示迭代过程,包含中心的变化和边界的变化

About

matlab project for Digital Image Processing, use SLIC super pixel segmentation and maximum flow graph cut to build an interactive fore/background segmentation app

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Languages