Globally optimal and region scalable CV model on automatic target segmentation
文献类型：会议论文
| 作者 | Fan HJ(范慧杰) ; Dong W(董威); Tang YD(唐延东)
|
| 出版日期 | 2010 |
| 会议名称 | 3rd International Conference on Intelligent Robotics and Applications, ICIRA 2010 |
| 会议日期 | November 10-12, 2010 |
| 会议地点 | Shanghai, China |
| 关键词 | Data handling Drop breakup Level measurement Numerical methods Optimization Robotics |
| 页码 | 600-608 |
| 中文摘要 | This paper proposes a new Globally Optimal and Region Scalable Chan-Vese model (GRCV model), which combines the advantages of both local and global intensity information. Intensity information in local regions is emphasized by adding a kernel function in the data fitting term, which thereby enables the proposed model to extract the fine structure in local region. The global information is also considered by adding one data fitting term in Piecewise Constant (PC) model to guarantee that proposed model can successfully segment complete target regions from intensity inhomogeneous images. We give the selection rules of combining the different terms under different target and background intensities: when the target is intensity inhomogeneous while the background is homogeneous, we choose the data fitting term which effects on the outside region of the evolution contour; conversely, we choose the other term. Experiments have been done on different images to compare the effectiveness of our methods with that of the classical PC model, Li's Local Binary Fitting (LBF) model, and Wang's Local and Global Intensity Fitting (LGIF) model. Comparison results show that our model obtains more satisfactory segmentation results. Moreover, it is robust to the curve initialization and noise in images. © 2010 Springer-Verlag. |
| 收录类别 | EI ; CPCI（ISTP） |
| 产权排序 | 1 |
| 会议录 | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
 |
| 会议录出版者 | Springer Verlag |
| 会议录出版地 | Heidelberg, Germany |
| 语种 | 英语 |
| ISSN号 | 0302-9743 |
| ISBN号 | 978-3-642-16583-2 |
| WOS记录号 | WOS:000324576300058 |
| 源URL | [http://ir.sia.cn/handle/173321/8724]  |
| 专题 | 沈阳自动化研究所_机器人学研究室 |
| 推荐引用方式 GB/T 7714 | Fan HJ,Dong W,Tang YD. Globally optimal and region scalable CV model on automatic target segmentation[C]. 见:3rd International Conference on Intelligent Robotics and Applications, ICIRA 2010. Shanghai, China. November 10-12, 2010. |
入库方式： OAI收割
来源：沈阳自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。