摘要
为了实现图像的自动分割,解决随机游走算法中标记点的自动选取问题,提出了一种结合显著性、模式挖掘算法以及随机游走的自动分割算法。计算图像的显著图,初步确定感兴趣目标区域,通过模式挖掘算法为随机游走初步选取标记点;结合显著目标位置信息进一步筛选标记点,分类为前景和背景标记点;使用随机游走算法对输入图像进行分割。实验结果表明:所提算法可以自动地对图像进行较精确地分割,在图像批量分割处理中具有一定的应用价值。
In order to realize automatic segmentation of image and solve automatic selection problem of mark points in random walk algorithm,propose an automatic segmentation algorithm combinaed significance,pattern mining algorithms and random walk. Calculate saliency map of image,primarily determine area of interested target.Select preliminary mark points for the random walk using pattern mining algorithms; combine with significant target location information for further selection of mark points,divide mark points into foreground and background mark points. Segmentation of the input image is carried out using random walk algorithm. Experimental results show that the algorithm can automatically and accurately segment images and has a certain application value in image batch segmentation processing.
作者
茅正冲
韩毅
MAO Zheng-chong;HAN Yi(Key Laboratory of Advanced Process Control for Light Industry, School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, China)
出处
《传感器与微系统》
CSCD
2018年第6期142-145,共4页
Transducer and Microsystem Technologies
基金
国家自然科学基金资助项目(60973095)
江苏省自然科学基金资助项目(BK20131107)
关键词
图像分割
自动分割
随机游走
显著性检测
模式挖掘
image segmentation
automatic segmentation
random walk
saliency detection
pattern mining
作者简介
茅正冲(1964-),男,硕士,副教授,主要从事机器人视听觉识别的研究工作,E-mail:maozcandxia@163.com。;韩毅(1992-),男,硕士研究生,主要研究方向为控制工程及应用,E-mail:jamin_hon@163.com。