| 基于度信息的图像过渡区提取与分割 |
| |
| 引用本文： | 闫成新,桑农,张天序.基于度信息的图像过渡区提取与分割[J].华中科技大学学报(自然科学版),2004,32(10):1-3. |
| |
| 作者姓名： | 闫成新 桑农 张天序 |
| |
| 作者单位： | 华中科技大学,图像识别与人工智能研究所,湖北,武汉,430074;华中科技大学,图像识别与人工智能研究所,湖北,武汉,430074;华中科技大学,图像识别与人工智能研究所,湖北,武汉,430074 |
| |
| 基金项目： | 国家自然科学基金重点资助项目 (6 0 135 0 2 0 ) |
| |
| 摘 要： | 提出一种基于度信息的图像过渡区提取与分割算法．将一幅图像映射为带权无向图，权函数定义为节点之间的相似度，由于目标内部与背景内部的像素同质性好，相似度大，因此度值较大，而边缘像素的度值就小，由此提取过渡区并进行图像分割．实验结果表明，本方法能够稳定地提取过渡区，且处理速度明显优于原有的过渡区提取方法。
|
| 关 键 词： | 图像分割 图论 过渡区 阈值 |
| 文章编号： | 1671-4512(2004)10-0001-03 |
| 修稿时间： | 2003年10月13 |
Degree-based image transition region extraction and its segmentation |
| |
| Yan Chengxin Sang Nong Zhang TianxuDoctoral candidate, Institute for Pattern Recognition & AI,Huazhong Univ. of Sci. & Tech.,Wuhan ,China..Degree-based image transition region extraction and its segmentation[J].JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE,2004,32(10):1-3. |
| |
| Authors: | Yan Chengxin Sang Nong Zhang TianxuDoctoral candidate Institute for Pattern Recognition & AI Huazhong Univ of Sci & Tech Wuhan China |
| |
| Institution: | Yan Chengxin Sang Nong Zhang TianxuDoctoral candidate, Institute for Pattern Recognition & AI,Huazhong Univ. of Sci. & Tech.,Wuhan 430074,China. |
| |
| Abstract: | A degree-based transition region extraction method (D-TREM) was proposed. If an image was converted into a weighted undirected graph, and the similarity between nodes the weight function defined, the degree of pixels inside the object and the background may be larger than those within the transition region. These properties of degree lead to our D-TREM. The experimental results demonstrated that D-TREM could steadily extract the transition region and the speed was significantly faster compared with other TRE method. |
| |
| Keywords: | image segmentation graph theory transition region threshold |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
