摘要
Snake模型(活动轮廓模型)大量应用于各种医学图像的分割,用于对超声图像病灶的分割和识别,可以大大提高临床诊断和决策的效率。针对NBGVF模型对弱边界处理效果差,分割运算量大,对初始轮廓不能自适应生成等不足,提出一种基于能量函数和颜色特征的初始轮廓提取方法,结合病灶区域颜色空间特征和医生先验信息,利用能量函数构建可变的Snake初始轮廓完成逼近,对图像中病灶位置进行有效的提取。实验结果表明,改进的方法提高了病灶分割和临床决策效率。
Recently, we focus more about special medical image segmentation and detection. The segmentation on Ultrasound image focus plays an important role in deciding. Snake model is widely used in the medical image segmentation. The model has been constantly improved and developed. We describe the Snake model on color liver ultrasoune image. Furthermore, the method is proposed that we use energy function and color information to detect the initial edge with the help of space detail and experience of doctor. Finally, the modified GVF Model is also used. The result shows that the algorithm is effective and simple.
出处
《井冈山大学学报（自然科学版）》
2014年第4期48-52,共5页
Journal of Jinggangshan University (Natural Science)
