集成Hough球面检测的3D非刚性ICP配准算法在脑肿瘤良恶性识别中的应用

Application of Brain Tumor Classification Based on Non-Rigid ICP Registration Employed with Hough Transform Sphere Detection

目前,脑部肿瘤已经成为世界上最致命的疾病之一。术前利用影像诊断手段对肿瘤良恶性检测的结果将会直接用于手术计划和实施方案的制定。Hough变换对脑部球面进行检测后,球心可作为固定对应点用于ICP配准点的约束,然后在非刚性配准的基础上用病患脑部图像与正常脑部图像进行3D差分,即可在差分区域的基础上进行特征提取和分类,用于脑肿瘤的良恶性识别。最终实验证明,该方法具有良好的识别分类效果。

Brain tumors have become one of the world's deadliest diseases.Preoperative detection of benign and malignant tumors by means of imaging diagnosis will directly affect the planning and implementation of surgery.Proposes a Hough transform which is used to analysis the brain region features with three points and to constraint the ICP.Then on the basis of non-rigid registration of patient brain images 3D finite difference with normal brain images,finally based on the regional feature extraction and classification,to identify the benign and ma.lignant brain tumors.The experiment demonstrates that the method has good recognition and classification effect.