中文摘要：

作为乳腺肿块检测的重要环节,肿块分割在乳腺癌的计算机辅助诊断系统中扮演着重要的角色。提出一种基于水平集的多尺度乳腺肿块分割方法。首先对乳腺图像进行高斯金字塔分解,在粗尺度图像上使用C-V模型对肿块进行粗分割,得到的粗轮廓作为细尺度图像上的初始轮廓;考虑到C-V模型在分割灰度不均匀图像时所存在的局限性,在细尺度上提出一种局部活动轮廓模型,对粗分割的结果进行局部精细化处理。另外,为了提高分割方法的自适应能力,从粗分割结果中抽取灰度、面积特征作为局部活动轮廓模型参数设定的依据。将本文方法、C-V模型以及RSF模型应用于89例肿块病灶图像时,分别获得0.236 1、0.300 4和0.373 8的平均误分率。结果表明,所提出的多尺度方法具有更高的分割精度和鲁棒性。

英文摘要：

As an important step of breast mass detection,mass segmentation plays an important role in many computer-aided diagnosis（CAD） systems for mammography.In this paper,we propose a novel scheme for mass segmentation in mammography,which is based on level set method and multi-scale analysis.Mammography is firstly decomposed into a sequence of images from fine to coarse using Gaussian pyramid,the C-V model is then applied at coarse scale,and the obtained rough contour is used as the initial contour for segmentation at fine scale.A local active contour model based on the image local information is utilized to refine the rough contour locally at fine scale and prevent the defect of the C-V model when it is applied on the images with asymmetric gray distribution.In order to improve the adaptivity of our method,the features of area and gray level extracted from the segmentation result at coarse scale is used for the parameter setting of the local active contour model.The proposed method,C-V model and RSF model were simultaneously applied for mass segmentation on 89 mammographic ROIs,and the achieved average ratios of misclassification error are 0.236 1,0.300 4 and 0.373 8,respectively.The result demonstrates that the proposed multi-scale segmentation method achieves a better performance in accuracy and robustness.