Hierarchical Non-linear Image Registration Integrating Deformable Segmentation

IntroductionSegmentation and registration are two impor-tant problems in the field of medical image analy-sis. Traditionally, solutions have been developedfor each of these two problems in relative isolationfrom the other, but with increasing dependence onthe existence of a solution for the other. For exam-ple, model-based segmentation methods[1,2]need toregister individual data sets to a common frame ofreference, so that statistics about the models canbe gathered to guide the evolution of the…     

基于点特征和边缘特征的无人机影像配准方法

为解决变形较大的无人机影像配准问题,提出了点特征和边缘特征相结合的配准方法.用尺度不变特征变换(SIFT)算法提取点特征,完成影像的初步配准,并通过多项式函数对影像进行粗校正.在此基础上提取影像的边缘特征信息,根据距离相似性对边缘特征信息进行配准;依据色彩能量差筛选点特征信息配准结果和边缘特征信息配准结果,采用小面元微分校正的方法对变形影像进行校正.实验结果表明:提出的配准方法能够弥补点特征配准方法和边缘特征配准方法的不足,其配准的鲁棒性提高10%左右,可以较好地完成变形较大的无人机影像配准.      

Mutual information optimization based dynamic log-polar image registration

Log-polar transformation(LPT)is widely used in image registration due to its scale and rotation invariant properties.Through LPT,rotation and scale transformation can be made into translation displacement in log-polar coordinates,and phase correlation technique can be used to get the displacement.In LPT based image registration,constant samples in digitalization processing produce less precise and effective results.Thus,dynamic log-polar transformation(DLPT)is used in this paper.DLPT is a method that generates several sample sets in axes to produce several results and only the effective results are used to get the final results by using statistical approach.Therefore,DLPT can get more precise and effective transformation results than the conventional LPT.Mutual information(MI)is a similarity measure to align two images and has been used in image registration for a long time.An optimal transform for image registration can be obtained by maximizing MI between the two images.Image registration based on MI is robust in noisy,occlusion and illumination changing circumstance.In this paper,we study image registration using MI and DLPT.Experiments with digitalizing images and with real image datasets are performed,and the experimental results show that the combination of MI with DLPT is an effective and precise method for image registration.     

MUTUAL INFORMATION BASED 3D NON-RIGID REGISTRATION OF CT/MR ABDOMEN IMAGES

Introduction　　 Automating medical image registration is be-coming an increasingly important goal. It formsthe basis of image processing techniques to com-bine information from the different modalities,which are now frequently acquired for many pa-tients. Such combined representations of patientimages have been shown to assistin the interpreta-tion of the complementary information provided bythe modalities[1,2 ] .　　 A variety of imaging modalities gives insightto different parameters and in…     

基于贝叶斯决策的航空影像自动配准

为消除像点投影差对航空影像目标监测中影像配准的影响,提出了一种基于贝叶斯决策理论的像点投影差消除方法.该方法通过推导分析像点投影差的分布规律及其对影像配准的影响,设置训练样本区,提取大量特征匹配对,计算得到投影差分布参数,并基于贝叶斯决策理论构建了像点投影差消除准则.为验证其有效性,选用KIT AIS影像数据进行了实验,比较分析了投影差消除前后的配准影像结果.分析结果显示,消除后差分视觉效果得到改善,信息熵减小10%,表明本文方法可有效提高航空影像配准的精度.      

Deformable Registration of MR Brains with Best Features

Introduction Deformable registration is very important formedical image analysis and so far various methodshave been proposed in decades[1-4]. HAMMER reg-istration algorithm[5]defines an attribute vector in-cluding intensity, edge type, and geometric mo-ment invariants as a signature of each point, to re-duce ambiguity in correspondence matching duringthe image registration. Another characteristic ofHAMMER is the hierarchical matching mecha-nism, which helps avoid the warping being trapped…     

基于一面两孔特征的点云配准方法

为了提高零件在扫描检测过程中点云与设计模型的配准精度,提出了一种基于一面两孔特征的点云配准方法.该方法粗配准以零件的平面/圆柱孔特征为对象,使设计模型和点云的局部坐标系重合,并通过改进ICP算法求解点云与设计模型最近点的距离最小平方和实现精配准.由于配准区域和最近点的计算方法不同,精配准进一步分为全域和特征域配准两种类型.全域精配准以距点云最近的设计模型三角网格点或投影点为最近点,适合于毛坯件;特征域精配准则通过求解点云在平面/圆柱孔特征上的投影点为最近点,适合于成品件.试验及计算结果表明:全域配准的配准精度随表面离散点距离的减小而提高, 当离散点距离达到1.50 mm时,其配准精度已经达到0.15 mm,基本满足工程应用要求.当配准精度相同时,配准效率较其它方法提高10%~20%.      

点云数据配准算法的研究

点云配准是点云数据处理的关键,直接影响最后合成结果和模型精度。目前,点云配准方法普遍存在对配准数据初始位姿要求高的缺点。将点云配准分为两个阶段：第一阶段是基于同名点配准,即粗配准,采用人机交互式,配准过程耗时短,节约时间;第二阶段是精配准,在粗配准后,依据最小二乘原理,用间接平差思想,通过最近点迭代算法对点云数据快速配准,并采用目标点集中、目标点坐标与转入目标点集中的点坐标中误差为指标,评价配准精度。进行粗配准的精配准不仅速度快、耗时短,并且可以避免因局部收敛而带来的局部最小问题。试验表明该方法有效可行。     

基于小波变换和线性判别分析的人脸识别方法

提出了将人脸图像的小波分解和线性判别分析结合以达到人脸识别的方法.首先对人脸图像作小波分解,并将分解后的低频系数进行线性判别分析进一步降低人脸特征向量的维数,最后利用最近邻分类器进行分类识别.实验表明,该方法的正确识别率高于传统的特征脸识别方法.     

Kinect-based automatic spatial registration framework for neurosurgical navigation

As image-guided navigation plays an important the pre-operative images with the intra-operative patient role in neurosurgery, the spatial registration mapping position becomes crucial for a high accurate surgi- cal output. Conventional landmark-based registration requires expensive and time-consuming logistic support. Surface-based registration is a plausible alternative due to its simplicity and efficacy. In this paper, we propose a comprehensive framework for surface-based registration in neurosurgical navigation, where Kinect is used to auto- matically acquire patient＇s facial surface in a real time manner. Coherent point drift （CPD） algorithm is employed to register the facial surface with pre-operative images （e.g., computed tomography （CT） or magnetic resonance imaging （MRI）） using a coarse-to-fine scheme. The spatial registration results of 6 volunteers demonstrate that the proposed framework has potential for clinical use.     

Groupwise Registration of Brain Magnetic Resonance Images:A Review

A relatively new family of image registration methods, namely groupwise registration, has recently emerged and been widely investigated due to its fundamentally key role in analyzing image populations in terms of atlas-based analysis or clinical diagnostic systems. Compared with pairwise registration, groupwise registration is capable of handling a large-scale population of images simultaneously in an unbiased way. In this paper, a review of the latest research on groupwise registration is presented. First, the schemes of pairwise registration and groupwise registration are compared. Then, a classification of groupwise registration and several exemplar implementations of groupwise registration are illustrated, including their experimental results. Finally, typical applications of groupwise registration, e.g., infant atlas construction, and population-based anatomical variability evaluation, are discussed in this paper.     

基于结构性字典学习的高光谱遥感图像分类

为提高高光谱遥感图像的分类精度,提出了一种新的结构性稀疏表示及字典学习的高光谱遥感图像分类方法.该方法能同时利用高光谱遥感图像像素间的空间及光谱关系得到表示每个像素的字典,被划分为同一像素组的像素具有通用的稀疏模式;由字典计算图像的稀疏表示系数获得遥感图像的稀疏表示特征;利用线性支持向量机算法实现对高光谱遥感图像的分类.对AVIRIS和ROSIS高光谱遥感图像进行的实验结果表明:提出的方法比普通字典学习分类精度分别提高0.041 1和0.046 6,Kappa系数分别提高0.179 3和0.056 3.      

基于互信息快速算法的多模医学图像配准

互信息作为相似性测度在多模医学图像配准领域得到了广泛应用,具有高精度和稳健的特点.但频繁的互信息计算降低了配准效率,同时对两幅图像重叠区域比较敏感.采用归一化互信息为测度,降低互信息计算中灰度级的快速算法,通过多模CT/MRI配准实验证明,可以在确保配准精度的同时缩短时间提高配准效率.     

基于NSCT变换的红外与可见光图像融合方法

针对小波变换不能很好地表达图像边缘信息，NSCT变换对图像细节信息表达缺失的问题，本文提出了一种改进的基于NSCT变换的图像融合方法．首先将经过预处理和配准后的红外图像和可见光图像进行NSCT变换，得到各个源图像的低频和高频系数，然后对分解后的低频系数采用小波变换的融合规则进行融合处理，高频系数则采用基于特征的区域能量的融合规则进行融合处理，最后对融合后的系数进行NSCT反变换得到融合图像．仿真实验表明，采用改进的NSCT融合方法对红外与可见光图像的融合有良好的效果，图像更清晰，信息更全面．     

基于LLE和LS_SVM的胃粘膜肿瘤细胞图像分类

胃粘膜肿瘤细胞图像的复杂性,组织器官形状的不规则性以及不同细胞的差异性,使得采用一般的线性分类方法对其进行分类很困难,结合局部线性嵌入（LLE）在处理非线性数据及最小二乘支持向量机（LS_SVM）在处理小样本、高维数及泛化问题方面的优势,文章提出一种基于LLE＋LS_SVM的胃粘膜肿瘤细胞图像分类方法,并采用LS_SVM的线性拟合误差来判断实验效果,最后比较本文方法与其他分类方法的优越性。实验结果表明,该方法在分类准确率和运行时间方面都有很大的优势。     

基于Parks-McClellan算法的UWB脉冲设计方法

针对UWB（ultra-wide band）脉冲波形的特点，提出了基于Parks-McClellan算法的UWB无线脉冲波形的优化设计方法，其基本思想是将UWB无线脉冲波形的优化设计等效为FIR（finite impulse response）滤波器的优化设计问题．对单周脉冲采用TH（time hopping）和二进制PPM（pulse position modulation）处理后得到的单频带和多频带模式UWB脉冲波形可充分满足FCC（federal communication committee）的频谱要求，相应的结果适用于单频带和多频带模式UWB系统．     

航空影像的几何校正研究与实现

航空侦察在军事上有着重要的地位,文中研究了航空影像的斜像校正.利用GPS实地采集地面控制点,采用多项式和双线性插值法对机上采集的图像进行了几何校正.校正后的图像可用于图像配准、融合和目标识别.     

节能型双频段无线视频监控系统

针对市场上对兼容多频段无线监控系统的需求,设计并实现了一套节能型双频段无线监控系统,并对其中的关键技术新型结构的带通滤波器和混合型光伏发电系统实现方式进行了论述分析,给出了在工程中实用的设计方法,其测试结果显示该系统可有效的应用于无线视频监控中.     

人脸非线性视角流形建模方法

针对线性子空间不足以描述头部视角空间非线性变化等因素影响人脸视角流形的精确建模问题,提出一种新的视角流形建模方法,并从理论上将该方法与经典的流形学习建模方法及概念驱动的视角流形建模方法进行比较,通过基于非线性张量分解的人脸及视角识别实验比较视角流形对识别结果的影响,从而给出视角流形的有效性比较.实验结果表明,本文提出的视角流形建模方法比概念驱动的视角流形和TensorFace中的线性视角系数均有更好的识别效果.     

基于高光谱成像技术的绝缘子污秽度预测

高光谱成像技术能对绝缘子进行非接触式成像,且具有多波段、图谱合一等特点. 为此,本文提出一种基于高光谱成像技术的绝缘子污秽度预测方法. 首先,利用高光谱成像仪对绝缘子进行成像,得到400～1 000 nm波段范围内的高光谱图像数据,并进行黑白校正;然后,获取感兴趣区域（region of interest,ROI）的反射率光谱曲线,进行Savitzky-Golay平滑、对数或一阶导数变换的预处理. 最后,联合部分的真实样本标签数据分别建立基于支持向量机的绝缘子污秽度预测（support vector machines-insulator contamination degree prediction,SVM-ICDP）和基于偏最小二乘回归的绝缘子污秽度预测（partial least squares regression-insulator contamination degree prediction,PLSR-ICDP）模型. 从实验结果中可知,当预处理方法采用一阶导数变换时,所建立的绝缘子污秽度预测模型效果最佳,即SVM-ICDP模型准确率达到91.84%;PLSR-ICDP模型的均方根误差（root mean square error,RMSE）为0.024 1.