摘要
目的:提出一种联合脉冲耦合神经网络改进模型(modified pulse coupled neural network,MPCNN)和多分辨奇异值分解(multi-resolution singular value decomposition,MSVD)的多模态医学图像融合算法。方法:第一步,采用MSVD将已配准的MRI和CT图像分解成高频和低频子图像;第二步,采用基于自适应连接因子的MPCNN方法融合低频系数,高频系数采用绝对值取大进行融合,最大限度保存图像细节信息;第三步,采用MSVD逆变换重建融合图像。结果:8组CT和MRI图像融合实验表明,基于提出算法获得的融合图像对比度、清晰度和边缘强度均最佳,且定量评价指标标准差、熵、互信息和边缘强度均高于其他融合算法。结论:提出的MPCNN算法能有效克服传统PCNN算法的局限性,与MSVD结合后融合性能优越,具有较高普适性和实用性,是一种可行的CT和MRI图像融合算法。
Objective: We proposed a novel multimodal medical images fusion method based on multi-resolution singular value decomposition(MSVD) and modified pulse coupled neural network(MPCNN). Methods: Firstly, the input pre-registered MRI and CT images are decomposed into high frequency(HF) and low frequency(LF) sub-bands by the MSVD transform. Then, the MPCNN model is applied adaptively to determine the linking strength. After that, LF coefficients are combined based on the output of MPCNN coefficients while HF coefficients are fused by the maximum selection rule. Finally, the inverse MSVD is applied to reconstruct the fused image. Results: Eight groups of multimodal MRI and CT images were used in simulation experiments. Visual analysis showed that the fused images of proposed method had better contrast, definition and edge intensity. Quantitative assessment showed that the standard deviation, entropy, mutual information and QAB/F in proposed method were superior to other methods in all groups. Conclusion: The MPCNN can overcome limits of conventional PCNN, when it combined with MSVD, the proposed method showed better robustness, superiority and become a feasible CT and MRI image fusion algorithm.
作者
宋方奔
缪正飞
张子齐
SONG Fang-ben;MIAO Zheng-fei;ZHANG Zi-qi(Department of Radiology, Nanjing First Hospital, Nanjing Medical University, Nanjing 210006, Jiangsu Province, P.R.C.)
出处
《中国数字医学》
2019年第7期9-12,共4页
China Digital Medicine
关键词
脉冲耦合神经网络
多分辨奇异值分解
多模态
医学图像融合
pulse coupled neural network
multi-resolution singular value decomposition
multimodal
medical image fusion
作者简介
通信作者:张子齐,南京医科大学附属南京医院(南京市第一医院)放射科,210006,江苏省南京市秦淮区长乐路68号.
