基于马尔可夫模型的多分辨率图像分割算法

Multi-resolution Image Segmentation Algorithm Based on Markov Model

为解决图像分割中过分割、欠分割和依赖初始分割问题,提出一种基于马尔可夫模型的多分辨率图像分割算法。利用变权重方法改进多分辨率马尔可夫随机场算法,结合曲波和小波变换对图像进行多分辨率分析,并通过区域合并减少图像中的区域数。实验结果表明,与经典算法相比,该算法的分割性能较好。

To solve the problem such as over-segmentation, under segmentation and dependence on initial segmented in image segmentation, this paper proposes a new multi-resolution image segmentation algorithm based on Markov model. It makes improvements on variable weight of Multi-resolution Markov Random Field（MRMRF） algorithm, combines curvelet transform and wavelet transform to multi-resolution analysis. It reduces area number of image by area merging. Experimental result proves the effectiveness of this algorithm compared with classical algorithms.