| 摘要: |
| 图像分割可视为两类模式分类问题.将最大似然聚类方法应用于图像分割,并采用模拟退火技术求解最大似然聚类,解决了用迭代方法求解最大似然聚类只能得到局部最优解的问题.获得的图像分割效果优于迭代方法和著名的Otsu方法,且分类误差小于迭代方法. |
| 关键词: 模拟退火 最大似然聚类 图像分割 |
| DOI： |
| 分类号: |
| 基金项目:国家自然科学基金资助项目(69803009) |
|
| Simulated Annealing based Maximum Likelihood Clustering Algorithm for Image Segmentation |
|
ZHANG Yin,PAN Yun-he
|
| Abstract: |
| Image segmentation can be regarded as the problem of two-class pattern classification. How to apply the maximum likelihood clustering algorithm to image segmentation is discussed in this paper. Simulated annealing technology is used to solve the problem of maximum likelihood clustering, which avoids the local optimal solution of iterative method. It shows better image segmentation effect than the famous Otsu algorithm and iterative method with less classification error than iterative method. |
| Key words: simulated annealing maximum likelihood clustering image segmentation |
