基于改进Snake模型的超声乳腺肿瘤轮廓提取

二维超声影像中肿瘤轮廓特征是判断乳腺肿瘤的良恶性的重要依据。针对超声医学图像的特点,本研究对经典的Snake模型进行了改进:内部能量中加入轮廓平均长度项的控制;外部能量由基于图像统计特征的区域能量以及梯度方向势能决定,并提出了基于贪婪算法求解模型最小值的快速算法。实验结果显示本算法在噪声强度较大的模拟图像和超声医学图像中均取得了同人工分割近似的结果,而经典的Snake模型和GVF模型受噪声干扰较大。大量的实验证明本算法有效地克服了散斑噪声对分割结果的影响,可准确高效地提取超声图像中的乳腺肿瘤轮廓。

Contour Extraction of Ultrasonic Breast Tumor Based on Modifed Snake Model

The contour of breast tumor in B-mode ultrasonic image is important indicator for discriminating the malignant tumor from those benign.The paper proposed a modification to the classical Snake model:first,a term of average contour length was added into the internal energy;second,the external energy is determined by the gradient directional energy together with the region energy which is based on regional statistic characteristics and insensitive to speckle noise;Finally,the greedy algorithm is used to find the minimum energy of the model.The modified model was tested by both synthetic image and real ultrasonic image.Comparing to the classical snake model and the gradient vector flow(GVF) model,the segmentation results of the proposed model is robust to the noise and approximate to the manual delineation results.Large numbers of experimental results show that the proposed model is accurate,fast and effective.