基于改进归一化割甲状腺肿瘤b超图像分割-ultrasound image segmentation of thyroid tumors based on improved normalized thyroidectomy.docx

摘要摘要现在甲状腺结节的发病率很高，这可能预示着甲状腺癌的出现。因此，准确判断出 甲状腺结节位置以及正确判别甲状腺结节是良性的还是恶性的是非常必要的。超声诊断 甲状腺结节的准确率不断提高，已成为甲状腺结节性疾病的首选辅助检查，但是超声的 回声、阴影、反射等特性也会使超声图像的质量退化，这就很难将其边缘精确地分割出。基于图论的图像分割技术是目前的研究热点，它通过定义一个采用图像某种割值来 表示目标函数，通过求解目标函数的极值实现分割。归一化割(Ncut)是其中的一种，它 主要是通过方向能量模型得到不同尺度子图像的边缘方向能量，进而得到干涉轮廓模型 的相似度，结合根据纹理基元得到图像纹理特征相似度，进而可以建立图像像素之间的 相似度，形成基于图像像素点的权值矩阵，构造无向加权图，最后运用特征值和特征向 量对图像进行分割。针对甲状腺肿瘤的 B 超图像在形成过程中产生的斑点噪声使图像质量变差的特点 以及归一化割(Ncut)方法对噪声敏感和计算复杂度高的缺点，本文将各向异性扩散模型 与归一化割相结合，对归一化割中的干涉轮廓线和权值矩阵作出了改进。所以图像经过 各向异性扩散模型增强处理去除了噪声，同时保留了重要的边缘和局部细节，使得改进 后的归一化割处理图像时减少了干涉轮廓线，在求取权值矩阵上降低了计算量，缩短了 分割时间，提高了分割结果的准确度。实验结果证明了该方法的可行性。关键词甲状腺超声图像归一化割各向异性扩散模型方向能量干涉轮廓线AbstractAbstractNow the incidence of thyroid nodules is high, it indicates the emergence of the thyroid cancer. Therefore, it is very necessary that accurately judge the location of thyroid nodules and correctly discriminate the enign and malignant of thyroid nodule . Ultrasonic diagnosis of thyroid nodules becomes the first auxiliary examination of thyroid nodular disease, but the characteristics of ultrasonic on echo, shadow, reflection and so on can make quality of the ultrasonic image degradation, so which is very difficult to part edge of the image precisely.Image segmentation technique based on graph theory is a hot spot in the present, it indicates the objective function by defining some cut values of image and segments images by solving the objective function extremum. Normalized cut is one of them, and it gets edge orientation energy of different scale subimage through the orientation energy model, then further to get similarity based on interference contour model, combined image texture feature similarity based on textons, which creates the similarity between image pixels. then forming weight matrix based on image pixel and structuring weighted graph, finally the algorithm segments images using the eigenvalue and eigenvector.In the processing of formating ultrasound for the thyroid tumour images, the characteristic which the speckle noise makes