Image registration methods in high‐dimensional space

Image matching using alpha-entropy measures and entropic graphs

We propose a novel multi-atlas segmentation method that employs a group-wise image registration method for the brain segmentation on rodent magnetic resonance (MR) images. The core element of the proposed segmentation is the use of a particle-guided image registration method that extends the concept of particle correspondence into the volumetric image domain. The registration method performs a group-wise image registration that simultaneously registers a set of images toward the space defined by the average of particles. The particle-guided image registration method is robust with low signal-to-noise ratio images as well as differing sizes and shapes observed in the developing rodent brain. Also, the use of an implicit common reference frame can prevent potential bias induced by the use of a single template in the segmentation process. We show that the use of a particle guided-image registration method can be naturally extended to a novel multi-atlas segmentation method and improves the registration method to explicitly use the provided template labels as an additional constraint. In the experiment, we show that our segmentation algorithm provides more accuracy with multi-atlas label fusion and stability against pair-wise image registration. The comparison with previous group-wise registration method is provided as well.

Registration Strategies and Similarity Measures for Three-dimensional Ultrasound Mosaicing

Attenuation correction of PET activation studies in the presenceof task-related motion

Summary form only given. In the manifold learning problem one seeks to discover a smooth low dimensional surface, i.e., a manifold embedded in a higher dimensional linear vector space, based on a set of measured sample points on the surface. In this paper we consider the closely related problem of estimating the manifold's intrinsic dimension and the intrinsic entropy of the sample points. Specifically, we view the sample points as realizations of an unknown multivariate density supported on an unknown smooth manifold. We present a novel geometrical probability approach, called the geodesic entropic graph (GET) method, to obtaining asymptotically consistent estimates of the manifold dimension and the Renyi /spl alpha/-entropy of the sample density on the manifold. The GET approach is striking in its simplicity and does not require reconstructing the manifold or estimating the multivariate density of the samples. The GET method simply constructs an entropic graph, e.g., a minimal spanning tree (MST) or k-nearest neighbor graph (k-NNG), to estimate the geodesic neighborhoods connecting points on the manifold. The growth rate of the length functional of the entropic graph is then used to simultaneously estimate manifold dimension and sample entropy.

The performances of information-theoretic multi-modality image registration methods crucially depend on the model representing the joint density function of the co-occurring image intensities and on the implementation of the entropy estimator. We proposed an entropy estimator for image registration based on quad-tree (QT) that is essentially an entropic graph entropy estimator, but can be adapted to work as a plug-in entropy estimator. This duality was achieved by incorporating the Hilbert kernel density estimator. Results of 3-D rigid-body registration of multi-modal brain volumes indicate that the proposed methods achieve similar accuracies as the registration method based on minimal spanning tree (MST), but have a higher success rate and a higher capture range. Although the MST and QT have similar computational complexities, the QT-based methods had about 50% shorter registration times.

This paper presents a novel method for pair-wise range image registration, a backbone task in world modeling, parts inspection and manufacture, object recognition, pose estimation, robotic navigation, and reverse engineering. The method finds the most suitable homogeneous transformation matrix between two constructed range images to create a more complete 3D view of a scene. The proposed solution integrates a ray casting-based fitness estimation with a global optimization method called improved self-adaptive differential evolution. This method eliminates the fine registration steps of the well-known iterative closest point (ICP) algorithm used in previously proposed methods, and thus, is the first direct global registration algorithm. With its parallel implementation potential, the ray casting-based algorithm speeds up the fitness calculation for the global optimization method, which effectively exploits the search space to find the best transformation solution. The integration was successfully implemented in a parallel paradigm on a multi-core computer processor to solve a simultaneous 3D localization problem. The fast, accurate, and robust results show that the proposed algorithm significantly improves on the registration problem over state-of-the-art algorithms.

In this paper, a kernel-based nonlinear version of the oblique subspace projection (OBSP) operator is defined in terms of kernel functions. Input data are implicitly mapped into a high-dimensional kernel feature space by a nonlinear mapping, which is associated with a kernel function. The OBSP expression is then derived in the feature space, which is kernelized in terms of the kernel functions in order to avoid explicit computation in the high-dimensional feature space. The resulting kernelized OBSP algorithm is equivalent to a nonlinear OBSP in the original input space. Experimental results based on simulated hyperspectral data and real hyperspectral imagery shows that the kernel oblique subspace projection (KOBSP) outperforms the conventional OBSP.

Brain magnetic resonance (MR) image registration is an important part of brain MR image registration. Practically, there are many images needing to be registered at same time and the requirement for time cost is high. The current image registration methods register images one by one independently, so the methods are called static image registration methods. The methods are time consuming since they need to randomly initialize the relevant parameters before they start to register some image, so they can not meet the requirement of the practical applications for time cost. Through statistical research, we find the differences of the brain MR images are limited due to some reasons such as similar poses, similar brain structures and so on. So, we proposed one dynamic brain MR image registration algorithm combining inheritance idea and PSO. The algorithm can inherit the information of the former image's registration and use it to guide the current image registration to improve the stability of high precision of image registration and reduce time cost for image registration. The experimental results show that this algorithm can obtain higher stability of image registration, higher registration precision and lower time cost.

Purpose: To enhance the speed and accuracy of certain treatment plan in radiation therapy, a real‐time dose calculation strategy is performed based on Fast Fourier Transformation (FFT) and CT image registration. For this purpose, different registration methods are achieved by using Insight Segmentation and Registration Toolkit (ITK) in MATLAB. Method and materials: Our group developed an efficient method for kernel calculation for pencil beam convolution‐superposition based on FFT. In this paper, different image registration methods are developed in MATLAB by using ITK, such as rigid 3D, similarity 3D, Demons deformation and BSpline based Free Form. As ITK is written and used in a C++ template, smart pointer and generic structure, an open source wrapper MATITK is modified to facilitate the use of its image processing capabilities while working in the high‐level environment of MATLAB. The image registration methods are demonstrated on head and neck (H&N) CT image sets of 6 patients. The results of evaluation and visualization for the registration are presented. Results and discussion: Transformation similarity and computation complexity as evaluation criteria are compared between rigid and deformable registration methods. Judging from the mutual information histogram and mean square metric, the deformable registration will get more improvement on similarity than the rigid registration. Checkerboard composites of the reference and registered images are displayed. The computational time of deformable registration are 10∼20 times longer with the whole set of images because of the pre‐processing, complicated transformation matrix and more parameters. A similarity3D registration method is adopted for the image registration of H&N radiation therapy to trade off between registration effect and the time consumption. Conclusion: The image registration can greatly contribute to the automatic contour adjustment and dose deformation for adaptive radiation therapy. A GUI program will be designed for the customized deformation region to speed up deformable registration.

Different tasks in forensic anthropology require the use of three-dimensional models of forensic objects (skulls, bones, corpses, etc) captured by 3D range scanners. Since a whole object cannot be completely scanned with a single image, multiple scans from different views are needed to supply the information to construct the 3D model. Range image registration methods study the accurate integration of the different views acquired by range scanners, with pair-wise approaches progressively processing every adjacent pair of scanned views until reconstructing the whole 3D model of the object. Our proposal is based on the adaptation of our previous work (Cordon et al, IEEE Conference on Evolutionary Computation, pp 2738–2744, 2005 in Pattern Recognit Lett 27(11); 1191-1200, 2006) in order to apply the scatter search evolutionary algorithm to pair-wise image registration in forensic anthropology applications. To measure the performance of this adaptation, we design an experimental setup considering some of the most recent and accurate evolutionary techniques for the problem, as well as one skull from our Physical Anthropology Lab. Two additional volumes, commonly used in other pair-wise range IR contributions, have also been considered to complement the comparison of results among the proposals.

In the superresolution (SR) technique, as an inverse problem, the generative model forms low-resolution images from a high-resolution image. This process consists of three principal operations, namely, warping, blurring, and down-sampling. Unlike the traditional SR methods in which a discrete generative model is used, in this paper, a continuous generative model has been proposed for the SR technique. Rather than using a bilinear interpolation for warping operation, we use a Gaussian kernel for interpolation, as it has lower aliasing effect. Within the framework of the proposed continuous generative model, the three principal operations are combined into a unified operation and hence performed simultaneously. According to this combinational operation, a simultaneous image registration and SR method is proposed. Unlike previous simultaneous methods, we neither calculate the Jacobian matrix numerically (to increase the accuracy of calculation) nor derive it treating the three principal operations separately (to reduce the error propagation). We develop a new approach to derive the Jacobian matrix analytically by combining the three principal operations. The results show that our proposed method is more efficient than the recently proposed simultaneous methods.

Abstract. This paper looks into the latest achievement in the low-cost Unmanned Aerial Vehicle (UAV) technology in their capacity to map the semi-development areas. The objectives of this study are to establish a new methodology or a new algorithm in image registration during interior orientation process and to determine the accuracy of the photogrammetric products by using UAV images. Recently, UAV technology has been used in several applications such as mapping, agriculture and surveillance. The aim of this study is to scrutinize the usage of UAV to map the semi-development areas. The performance of the low cost UAV mapping study was established on a study area with two image processing methods so that the results could be comparable. A non-metric camera was attached at the bottom of UAV and it was used to capture images at both sites after it went through several calibration steps. Calibration processes were carried out to determine focal length, principal distance, radial lens distortion, tangential lens distortion and affinity. A new method in image registration for a non-metric camera is discussed in this paper as a part of new methodology of this study. This method used the UAV Global Positioning System (GPS) onboard to register the UAV image for interior orientation process. Check points were established randomly at both sites using rapid static Global Positioning System. Ground control points are used for exterior orientation process, and check point is used for accuracy assessment of photogrammetric product. All acquired images were processed in a photogrammetric software. Two methods of image registration were applied in this study, namely, GPS onboard registration and ground control point registration. Both registrations were processed by using photogrammetric software and the result is discussed. Two results were produced in this study, which are the digital orthophoto and the digital terrain model. These results were analyzed by using the root mean square errors and mean absolute error to determine the level of accuracy and the precision of photogrammetric products. It can be concluded that the new method of image registration by using the GPS onboard of the UAV produces medium accuracy result compared to the method that uses the ground control point. This new method can be used for the medium accuracy requirements. Unmanned Aerial Vehicle can be used for several applications, which requires a medium accuracy.

Automated methods for image segmentation, image registration, clustering of images and probabilistic atlas construction are of great interest in medical image analysis. In this work, we propose a model where these different aspects are combined in one comprehensive probabilistic framework. The framework is formulated as an EM optimization algorithm. Validation is performed on simulated and real images in terms of segmentation, clustering and atlas construction. Accurate segmentations are obtained and the different modes in a population of normal controls and Huntington disease patients are discovered. Furthermore, our method reveals the localization of cluster specific morphological differences for each image in the population.

Purpose: During a course of radiation therapy for the prostate, organs such as the bladder and rectum may change shape significantly from day to day, causing the delivered dose to differ from that planned. Deformable image registration has been used to overlap structures with different shape. In this project, we propose a method to validate the correspondence accuracy of a deformable registration algorithm using images of previously treated prostate patients. Method and Materials: The planning and one daily treatment CT images for each of five patients with three Calypso transponders implanted in their prostate were retrospectively selected for the study. Two CT volumes of each patient were pre‐processed to mask the fiducials. A fluid‐based image registration method was used to register images. The displacement fields obtained were applied on the original images with fiducials. The overlap of prostate volumes between the registered images was evaluated by the coincidence index (CI). The centroid positions of each transponder were calculated. The distance between the corresponding transponders was used to quantify the accuracy of correspondence of image registration. Results: The fluid deformable image registration produced satisfactory results of the overlap of the prostate volumes. The mean CI is 95.3%, with 91.5% as minimum and 98.7% as maximum. The average distance between the centroid of all pairs of transponders is 0.35 mm. The displacement is related to the quality of image registration. For image registration with CI 98.7%, all displacement of three pairs of transponders is within 0.22 mm. When less overlap is achieved, CI 91.5%, one pair of transponder is 0.78 mm apart. Conclusion: We have used real patient data to validate the correspondence of the fluid deformable image registration method. For the prostate, it has maintained sub‐millimeter mapping accuracy, which is dependent on image registration quality.