图像分割方法综述研究

图像分割是计算机视觉领域重要和基础性的问题，也是颇具挑战性的任务。为了解该问题的研究现状、存在问题及发展前景，在广泛调研现有文献和最新成果的基础上，针对2000年之后主流的图像分割方法进行了研究，将之分为四类：基于图论的方法、基于聚类的方法、基于分类的方法以及结合聚类和分类的方法，对每类方法所包含的典型算法，尤其是该领域最近几年发表的最新文章的基本思想、优缺点进行介绍和分析。最后介绍了图像分割常用的基准数据集和算法评价指标，对比各种算法并总结全文，对未来可能的发展趋势进行了展望。

Survey on Image Segmentation Methods

Image segmentation is an important and fundamental problem in computer vision, meanwhile it''s a challenging task. In order to find out the state-of-the-art, main problems and future trends of image segmentation, this paper introduces the mainstream image segmentation methods after 2000 on the basis of extensive research on the existing literatures and the latest achievements. These methods are categorized into four classes: graph theory based methods, clustering based methods, classification based methods, and hybrid methods of clustering and classification. The basic ideas, advantage and disadvantage of typical algorithms belong to each category, especially the most recently published papers are introduced and analyzed. Finally, this paper introduces the datasets which are commonly used as benchmark and evaluation metrics, compares all the algorithms, summarizes the work and forecasts some potential future research work.