基于Piella指数优化的CT和MR图像融合.docVIP

基于Piella指数优化的CT和MR图像融合余南南，邱天爽，毕峰，王爱齐 （大连理工大学电子信息与电气工程学部，大连 116024） 摘要：目的 实现CT与MR图像融合，从而帮助医生做出准确的诊断和治疗。方法 将Piella指数最大化作为图像融合的目标，首先通过滑动窗对配准后的原始图像分块，并将每一个图像块字典编纂为列向量；然后通过优化算法更新图像块加权向量，使Piella指数最大；最后根据加权向量合并相对应的图像块，并整理得到融合图像。结果 仿真实验对20组CT和MR图像进行融合处理，将本文算法与3种流行的融合算法比较，并给出量化评价指标，表明本文算法性能优越。结论 该算法为多模态医学图像融合提供了一种新的有效手段。 关键词：图像融合；Piella指数；CT；MR CT and MR Image Fusion based on Piella index, YU Nan-nan, QIU Tian-shuang, BI Feng, WANG Ai-qi. (Faculty of Electronic Information and Electrical Engineering, Dalian University of Technology, Dalian 116024) Abstract: Objective To fuse CT and MR images in order to provide doctors with accurate diagnosis and treatment. Method The maximization of the Piella index was used as object of image fusion. Firstly，the authors used the sliding window technique to divide each source image into small patches, and each patch was ordered lexicographically as vector. Secondly, through the optimization algorithm, the weights were determined to maximize the Piella index. Finally, the corresponding patches were combined according to the weights and the fused image is reconstructed by averaging the overlapped image patches. Result Computer simulations were conducted using twenty sets of CT and MR images. We compared our method with three state-of-the-art methods and also evaluated their fusion results through quality evaluation methods. The proposed scheme had better fusion performance. Conclusion This method provides a new useful tool for multi-modal images fusion. Key words: image fusion; Piella index; CT; MR Address reprint request to: QIU Tian-shuang. Faculty of Electronic Information and Electrical Engineering, Dalian University o f Technology, Dalian Liaoning 116024, China 近年来，医学成像技术广泛应用于疾病的诊断和治疗中，其中CT和MR图像是两个非常重要的医学图像。由于成像原理不同，它们侧重不同的人体信息。CT是根据人体组织对X射线的吸收率不同而成像，骨组织成像清晰；MR是根据在强磁场中射频脉冲和氢核的相互作用而获得的，对骨组织和钙化点信号较弱，但可清楚分辨出肌肉、筋膜、脂肪、脑灰质、脑白质等软组织。这两种图像具有很强的互补性，因此CT和MR图像融合对于疾病的诊断和治疗具有重要的意义[1~2]。图像融合包含两个部分：配准和融合。配准是指将通过不同手段得到的图像进行几何位置配准,即寻找某种空间变换,使两幅图像的对应点达到空间