【摘要】 提出了一种新的方法,用于解决数字人数据集重建前的切片图像配准和分割问题.首先利用改进的彩色空间动态阈值法和形态学处理提取切片图像的定位杆坐标,在此基础上对原始图像进行变换,从而实现数据集的配准;其次根据配准后数据集相邻切片目标轮廓在位置和形状上的相似性,利用前层切片分割结果指导初始化位置和限制搜索空间的主动轮廓模型,实现了配准后数据集的自动分割.所提出的方法在数字人切片图像的分割中取得了令人满意的效果,并且该方法还可以应用在CT、MRI等医学图像序列中的器官和部件的分割中.更多还原

【Abstract】 A new method was proposed to solve the registration and segmentation problem of slice images for digital human dataset reconstruction. First,locating rod was found through the improved auto-threshold method of color-space and morphology operation,and the origin datasets were transformed to implement auto-registration.Then,according to the similarity of position and shape of registered image sequence,an active contour model was involved whose initial contour and searching space were guided by the result of the previous slice to segment each slice automatically.The proposed method gets encouraging result and is also suitable for the segmentation of organs and parts from CT or MRI image series.更多还原