工程科学学报，第41卷，第7期：955-960.2019年7月 Chinese Journal of Engineering,Vol.41,No.7:955-960,July 2019 DOI:10.13374/j.issn2095-9389.2019.07.015;http://journals.ustb.edu.cn 一种改进的非刚性图像配准算法 何凯，魏颖，王阳，黄婉蓉 天津大学电气自动化与信息工程学院，天津300072 区通信作者，E-mail:hekai(@tju.edu.cn 摘要非刚性图像配准一直是计算机视觉领域的研究重点.为解决上述问题，提出一种改进的光流场模型算法，以提高光 流估计的准确度.算法首先对原始变分光流模型进行了改进，提出利用新的各向异性正则项来代替原来的同向扩散函数，以 避免图像模糊，保留图像的边缘特征与细节特征：此外，通过引入包含邻域信息的非局部平滑项来去除光流噪点，同时增加了 一个结合图像结构与光流运动信息的权函数，以减少过平滑所造成的细节丢失，提高算法的鲁棒性.最后，利用交替最小化与 金字塔分层迭代策略相结合的方法求解位移场，实现非刚性图像的自动配准.仿真实验结果表明，与传统方法相比，本文算法 对不同类型的非刚性图像均具有较高的鲁棒性，取得了理想的图像配准效果 关键词非刚性图像配准；光流场模型；各向异性正则项；非局部平滑项：交替最小化 分类号TP391.41 An improved non-rigid image registration approach HE Kai,WEI Ying,WANG Yang,HUANG Wan-rong School of Electrical and Information Engineering,Tianjin University,Tianjin 300072,China XCorresponding author,E-mail:hekai@tju.edu.cn ABSTRACT With the rapid development of image registration technology,it is being widely used in the fields of medical image pro- cessing,remote sensing image analysis,computer vision,and others.Image registration involves two or more images that contain the same object that are obtained under different conditions.Geometric mapping between images is realized by spatial geometric transforma- tion,so that the points in one image can be related to their corresponding points in the other.Compared with rigid transformations, non-rigid transformations usually have severe local distortions and obvious nonlinear characteristics.So,it is difficult to describe non- rigid transformations using a unified transformation model.For this reason,non-rigid image registration has always been an issue and a source of difficulty in the field of computer vision.To solve this problem,an improved optical-flow-model algorithm was proposed to more accurately estimate the optical flow field.First,the original variational optical flow model was improved.To prevent blurring and preserve the edge and detail features of images,a new anisotropic regular term was proposed to replace the original homologous diffusion term.Then,to remove optical flow outliers,a non-local smoothness term was introduced that contained neighborhood information.Mo- reover,a weight function that combines image-structure and optical-flow information was added to reduce the loss of detail caused by over-smoothing and to improve robustness.Finally,to solve the displacement field and realize the automatic registration of non-rigid im- ages,an alternating minimization method and pyramid hierarchical iteration strategy were utilized.To verify the effectiveness of the pro- posed algorithm,subjective and objective evaluation values such as the peak signal-to-noise ratio PSNR)and normalized mutual infor- mation (NMI)were adopted to analyze the registration results.Compared with state-of-the-art methods,experimental results reveal the robustness and ideal registration effects of the proposed method on different types of non-rigid images. KEY WORDS non-rigid image registration;optical flow model;anisotropic regularization term;non-local smoothness term;alternate minimization 收稿日期：2018-05-16 基金项目：国家自然科学基金资助项目(61271326)工程科学学报,第 41 卷,第 7 期:955鄄鄄960,2019 年 7 月 Chinese Journal of Engineering, Vol. 41, No. 7: 955鄄鄄960, July 2019 DOI: 10. 13374 / j. issn2095鄄鄄9389. 2019. 07. 015; http: / / journals. ustb. edu. cn 一种改进的非刚性图像配准算法 何 凯苣 , 魏 颖, 王 阳, 黄婉蓉 天津大学电气自动化与信息工程学院, 天津 300072 苣通信作者, E鄄mail: hekai@ tju. edu. cn 摘 要 非刚性图像配准一直是计算机视觉领域的研究重点. 为解决上述问题,提出一种改进的光流场模型算法,以提高光 流估计的准确度. 算法首先对原始变分光流模型进行了改进,提出利用新的各向异性正则项来代替原来的同向扩散函数,以 避免图像模糊,保留图像的边缘特征与细节特征;此外,通过引入包含邻域信息的非局部平滑项来去除光流噪点,同时增加了 一个结合图像结构与光流运动信息的权函数,以减少过平滑所造成的细节丢失,提高算法的鲁棒性. 最后,利用交替最小化与 金字塔分层迭代策略相结合的方法求解位移场,实现非刚性图像的自动配准. 仿真实验结果表明,与传统方法相比,本文算法 对不同类型的非刚性图像均具有较高的鲁棒性,取得了理想的图像配准效果. 关键词 非刚性图像配准; 光流场模型; 各向异性正则项; 非局部平滑项; 交替最小化 分类号 TP391郾 41 收稿日期: 2018鄄鄄05鄄鄄16 基金项目: 国家自然科学基金资助项目(61271326) An improved non鄄rigid image registration approach HE Kai 苣 , WEI Ying, WANG Yang, HUANG Wan鄄rong School of Electrical and Information Engineering, Tianjin University, Tianjin 300072, China 苣Corresponding author, E鄄mail: hekai@ tju. edu. cn ABSTRACT With the rapid development of image registration technology, it is being widely used in the fields of medical image pro鄄 cessing, remote sensing image analysis, computer vision, and others. Image registration involves two or more images that contain the same object that are obtained under different conditions. Geometric mapping between images is realized by spatial geometric transforma鄄 tion, so that the points in one image can be related to their corresponding points in the other. Compared with rigid transformations, non鄄rigid transformations usually have severe local distortions and obvious nonlinear characteristics. So, it is difficult to describe non鄄 rigid transformations using a unified transformation model. For this reason, non鄄rigid image registration has always been an issue and a source of difficulty in the field of computer vision. To solve this problem, an improved optical鄄flow鄄model algorithm was proposed to more accurately estimate the optical flow field. First, the original variational optical flow model was improved. To prevent blurring and preserve the edge and detail features of images, a new anisotropic regular term was proposed to replace the original homologous diffusion term. Then, to remove optical flow outliers, a non鄄local smoothness term was introduced that contained neighborhood information. Mo鄄 reover, a weight function that combines image鄄structure and optical鄄flow information was added to reduce the loss of detail caused by over鄄smoothing and to improve robustness. Finally, to solve the displacement field and realize the automatic registration of non鄄rigid im鄄 ages, an alternating minimization method and pyramid hierarchical iteration strategy were utilized. To verify the effectiveness of the pro鄄 posed algorithm, subjective and objective evaluation values such as the peak signal鄄to鄄noise ratio (PSNR) and normalized mutual infor鄄 mation (NMI) were adopted to analyze the registration results. Compared with state鄄of鄄the鄄art methods, experimental results reveal the robustness and ideal registration effects of the proposed method on different types of non鄄rigid images. KEY WORDS non鄄rigid image registration; optical flow model; anisotropic regularization term; non鄄local smoothness term; alternate minimization