基于高斯混合模型的有约束彩色医学图像群配准技术  

Ensemble registration of medical images with Gaussian mixture model and color component regularization

作　　者：王玉文[1] 胡顺波[2] 

机构地区：[1]临沂大学图书馆,山东临沂276005 [2]临沂大学信息学院,山东临沂276005

出　　处：《计算机应用》2014年第1期154-157,共4页journal of Computer Applications

基　　金：国家自然科学基金资助项目(61102040;31000744);山东省高校科研发展计划项目(J13LN85)

摘　　要：为利用多幅彩色图像中的丰富信息,提高配准精度和效果,把基于高斯混合模型的灰度图像群配准技术推广为彩色图像群配准技术。为了减少同一彩色图像不同彩色分量之间的形变偏差,定义了彩色分量形变约束项。总彩色图像配准测度是两个函数的代数和构造:彩色分量形变约束项和基于概率分布的对数似然函数。通过对人体胃部彩色图像和数据可视人彩色切片图像的群配准实验,证实了有约束彩色图像群配准技术具有很好的配准结果。In order to use the abundant information among several color images, and enhance the registration accuracy, the ensemble registration based on Gaussian Mixture Model （GMM） was extended from gray images to color images in this paper. To decrease the difference among color component deformations of the same image, the color component regularization term was incorporated into ensemble registration and a new total cost function was formulated. Color ensemble registration was applied to gastroseope images and tissue section images of color visible human. The test results show that the proposed color ensemble registration method can successfully align color images.

关 键 词：彩色图像配准 群配准 彩色分量约束 高斯混合模型 

分 类 号：TN911.73[电子电信—通信与信息系统]