基于活动轮廓的多分辨率自适应图像分割

摘要

本文在活动轮廓模型的基础上，提出了一种自适应图像分割方法.引入了新的与图象统计信息、梯度信息有关的加权外部能量，使得分割结果与模型的初始位置无关，不受噪声影响；利用ACD方法使模型自适应地改变其拓扑结构；为了提高图象分割的速度和鲁棒性，提出了多分辨率图象分割算法.利用该方法对一些形状、拓扑结构复杂的物体进行了分割实验，结果验证了该方法的有效性.%Adaptive image segmentation on active contours is presented inthis paper. We introduce a new external energy from the statistic characteristics of image and gradient of pixel intensities,which make segmentation independent on the initial active contour. The Affine Cell Decomposition method is used to adaptively change topologies of model by reparameter the active contour. In order to improve the speed and robust of segmetation, the muir-resolution image segmentation algorithm is proposed.Results of several experiments are shown demonstrating the potential of our method for segmenting some objects with complicated shape and topologies.