FMRI 3D registration based on Fourier space subsets using neural networks

In this work, we present a neural network (NN) based method designed for 3D rigid-body registration of FMRI time series, which relies on a limited number of Fourier coefficients of the images to be aligned. These coefficients, which are comprised in a small cubic neighborhood located at the first octant of a 3D Fourier space (including the DC component), are then fed into six NN during the learning stage. Each NN yields the estimates of a registration parameter. The proposed method was assessed for 3D rigid-body transformations, using DC neighborhoods of different sizes. The mean absolute registration errors are of approximately 0.030 mm in translations and 0.030 deg in rotations, for the typical motion amplitudes encountered in FMRI studies. The construction of the training set and the learning stage are fast requiring, respectively, 90 s and 1 to 12 s, depending on the number of input and hidden units of the NN. We believe that NN-based approaches to the problem of FMRI registration can be of great interest in the future. For instance, NN relying on limited K-space data (possibly in navigation echoes) can be a valid solution to the problem of prospective (in frame) FMRI registration.

Medical image registration based on random line sampling

One of the key aspects in 3D-image registration is the computation of the joint intensity histogram. We propose a new approach to compute this histogram using uniformly distributed random lines to sample stochastically the overlapping volume between two 3D-images. The intensity values are captured from the lines at evenly spaced positions, taking an initial random offset different for each line. This method provides us with an accurate, robust and fast mutual information-based registration. The interpolation effects are drastically reduced, due to the stochastic nature of the line generation, and the alignment process is also accelerated. The results obtained show a better performance of the introduced method than the classic computation of the joint histogram

Assessment of myocardial edema and area-at-risk in acute myocardial infarction by CMR: Evaluation of a novel T2-mapping method

Objectives. The goal of this study is to evaluate a T2-mapping sequence by: (i) measuring the reproducibility intra- and inter-observer variability in healthy volunteers in two separate scanning session with a T2 reference phantom; (2) measuring the mean T2 relaxation times by T2-mapping in infarcted myocardium in patients with subacute MI and compare it with patient's the gold standard X-ray coronary angiography and healthy volunteers results. Background. Myocardial edema is a consequence of an inflammation of the tissue, as seen in myocardial infarct (MI). It can be visualized by cardiovascular magnetic resonance (CMR) imaging using the T2 relaxation time. T2-mapping is a quantitative methodology that has the potential to address the limitation of the conventional T2-weighted (T2W) imaging. Methods. The T2-mapping protocol used for all MRI scans consisted in a radial gradient echo acquisition with a lung-liver navigator for free-breathing acquisition and affine image registration. Mid-basal short axis slices were acquired.T2-maps analyses: 2 observers semi- automatically segmented the left ventricle in 6 segments accordingly to the AHA standards. 8 healthy volunteers (age: 27 ± 4 years; 62.5% male) were scanned in 2 separate sessions. 17 patients (age : 61.9 ± 13.9 years; 82.4% male) with subacute STEMI (70.6%) and NSTEMI underwent a T2-mapping scanning session. Results. In healthy volunteers, the mean inter- and intra-observer variability over the entire short axis slice (segment 1 to 6) was 0.1 ms (95% confidence interval (CI): -0.4 to 0.5, p = 0.62) and 0.2 ms (95% CI: -2.8 to 3.2, p = 0.94, respectively. T2 relaxation time measurements with and without the correction of the phantom yielded an average difference of 3.0 ± 1.1 % and 3.1 ± 2.1 % (p = 0.828), respectively. In patients, the inter-observer variability in the entire short axis slice (S1-S6), was 0.3 ms (95% CI: -1.8 to 2.4, p = 0.85). Edema location as determined through the T2-mapping and the coronary artery occlusion as determined on X-ray coronary angiography correlated in 78.6%, but only in 60% in apical infarcts. All except one of the maximal T2 values in infarct patients were greater than the upper limit of the 95% confidence interval for normal myocardium. Conclusions. The T2-mapping methodology is accurate in detecting infarcted, i.e. edematous tissue in patients with subacute infarcts. This study further demonstrated that this T2-mapping technique is reproducible and robust enough to be used on a segmental basis for edema detection without the need of a phantom to yield a T2 correction factor. This new quantitative T2-mapping technique is promising and is likely to allow for serial follow-up studies in patients to improve our knowledge on infarct pathophysiology, on infarct healing, and for the assessment of novel treatment strategies for acute infarctions.

Dense Deformation Field Estimation for Atlas Registration using the Active Contour Framework

In this paper, we propose a new paradigm to carry outthe registration task with a dense deformation fieldderived from the optical flow model and the activecontour method. The proposed framework merges differenttasks such as segmentation, regularization, incorporationof prior knowledge and registration into a singleframework. The active contour model is at the core of ourframework even if it is used in a different way than thestandard approaches. Indeed, active contours are awell-known technique for image segmentation. Thistechnique consists in finding the curve which minimizesan energy functional designed to be minimal when thecurve has reached the object contours. That way, we getaccurate and smooth segmentation results. So far, theactive contour model has been used to segment objectslying in images from boundary-based, region-based orshape-based information. Our registration technique willprofit of all these families of active contours todetermine a dense deformation field defined on the wholeimage. A well-suited application of our model is theatlas registration in medical imaging which consists inautomatically delineating anatomical structures. Wepresent results on 2D synthetic images to show theperformances of our non rigid deformation field based ona natural registration term. We also present registrationresults on real 3D medical data with a large spaceoccupying tumor substantially deforming surroundingstructures, which constitutes a high challenging problem.

Study Regarding Implementation of a Uniform Statewide System to Allow for Electronic Transactions for the Registration and Titling of Motor Vehicles

LEGISLATIVE STUDY – The 83rd General Assembly of the Iowa Legislature, in Senate File 2273, directed the Iowa Department of Transportation (DOT) to conduct a study of how to implement a uniform statewide system to allow for electronic transactions for the registration and titling of motor vehicles. PARTICIPANTS IN STUDY – As directed by Senate File 2273, the DOT formed a working group to conduct the study that included representatives from the Consumer Protection Division of the Office of the Attorney General, the Department of Public Safety, the Department of Revenue, the Iowa State County Treasurer’s Association, the Iowa Automobile Dealers Association, and the Iowa Independent Automobile Dealers Association. CONDUCT OF THE STUDY – The working group met eight times between June 17, 2010, and October 1, 2010. The group discussed the costs and benefits of electronic titling from the perspectives of new and used motor vehicle dealers, county treasurers, the DOT, lending institutions, consumers and consumer protection, and law enforcement. Security concerns, legislative implications, and implementation timelines were also considered. In the course of the meetings the group: 1. Reviewed the specific goals of S.F. 2273, and viewed a demonstration of Iowa’s current vehicle registration and titling system so participants that were not users of the system could gain an understanding of its current functionality and capabilities. 2. Reviewed the results of a survey of county treasurers conducted by the DOT to determine the extent to which county treasurers had processing backlogs and the extent to which county treasurers limited the number of dealer registration and titling transactions that they would process in a single day and while the dealer waited. Only eight reported placing a limit on the number of dealer transactions that would be processed while the dealer waited (with the number ranging from one to four), and only 11 reported a backlog in processing registration and titling transactions as of June 11, 2010, with most backlogs being reported in the range of one to three days. 3. Conducted conference calls with representatives of the American Association of Motor Vehicle Administrators (AAMVA) and representatives of three states -- Kansas, which has an electronic lien and titling (ELT) program, and Wisconsin and Florida, each of which have both an ELT program and an electronic registration and titling (ERT) program – to assess current and best practices for electronic transactions. In addition, the DOT (through AAMVA) submitted a survey to all U.S. jurisdictions to determine how, if at all, other states implemented electronic transactions for the registration and titling of motor vehicles. Twenty-eight states responded to the survey; of the 28 states that responded, only 13 allowed liens to be added or released electronically, and only five indicated allowing applications for registration and titling to be submitted electronically. DOT staff also heard a presentation from South Dakota on its ERT system at an AAMVA regional meeting. ELT information that emerged suggests a multi-vendor approach, in which vendors that meet state specifications for participation are authorized to interface with the state’s system to serve as a portal between lenders and the state system, will facilitate electronic lien releases and additions by offering lenders more choices and the opportunity to use the same vendor in multiple states. The ERT information that emerged indicates a multi-interface approach that offers an interface with existing dealer management software (DMS) systems and through a separate internet site will facilitate ERT by offering access that meets a variety of business needs and models. In both instances, information that emerged indicates that, in the long-term, adoption rates are positively affected by making participation above a certain minimum threshold mandatory. 4. To assess and compare functions or services that might be offered by or through a vendor, the group heard presentations from vendors that offer products or services that facilitate some aspect of ELT or ERT. 5. To assess the concerns, needs and interest of Iowa motor vehicle dealers, the group surveyed dealers to assess registration and titling difficulties experienced by dealers, the types of DMS systems (if any) used by dealers, and the dealers’ interest and preference in using an electronic interface to submit applications for registration and titling. Overall, 40% of the dealers that responded indicated interest and 57% indicated no interest, but interest was pronounced among new car dealers (75% were interested) and dealers with a high number of monthly transactions (85% of dealers averaging more than 50 sales per month were interested). The majority of dealers responding to the dealer survey ranked delays in processing and problems with daily limits on transaction as ―minor difficulty or ―no difficulty. RECOMMENDATIONS -- At the conclusion of the meetings, the working group discussed possible approaches for implementation of electronic transactions in Iowa and reached a consensus that a phased implementation of electronic titling that addressed first electronic lien and title transactions (ELT) and electronic fund transfers (EFT), and then electronic applications for registration and titling (ERT) is recommended. The recommendation of a phased implementation is based upon recognition that aspects of ELT and EFT are foundational to ERT, and that ELT and EFT solutions are more readily and easily attained than the ERT solution, which will take longer and be somewhat more difficult to develop and will require federal approval of an electronic odometer statement to fully implement. ELT – A multi-vendor approach is proposed for ELT. No direct costs to the state, counties, consumers, or dealers are anticipated under this approach. The vendor charges participating lenders user or transaction fees for the service, and it appears the lenders typically absorb those costs due to the savings offered by ELT. Existing staff can complete the programming necessary to interface the state system with vendors’ systems. The estimated time to implement ELT is six to nine months. Mandatory participation is not recommended initially, but should be considered after ELT has been implemented and a suitable number of vendors have enrolled to provide a fair assessment of participation rates and opportunities. EFT – A previous attempt to implement ELT and EFT was terminated due to concern that it would negatively impact county revenues by reducing interest income earned on state funds collected by the county and held until the monthly transfer to the state. To avoid that problem in this implementation, the EFT solution should remain revenue neutral to the counties, by allowing fees submitted by EFT to be immediately directed to the proper county account. Because ARTS was designed and has the capacity to accommodate EFT, a vendor is not needed to implement EFT. The estimated time to implement EFT is six to nine months. It is expected that EFT development will overlap ELT development. ERT – ERT itself must be developed in phases. It will not be possible to quickly implement a fully functioning, paperless ERT system, because federal law requires that transfer of title be accompanied by a written odometer statement unless approval for an alternate electronic statement is granted by the National Highway Traffic Safety Administration (NHTSA). It is expected that it will take as much as a year or more to obtain NHTSA approval, and that NHTSA approval will require design of a system that requires the seller to electronically confirm the seller’s identity, make the required disclosure to the buyer, and then transfer the disclosure to the buyer, who must also electronically confirm the buyer’s identity and electronically review and accept the disclosure to complete and submit the transaction. Given the time that it will take to develop and gain approval for this solution, initial ERT implementation will focus on completing and submitting applications and issuing registration applied for cards electronically, with the understanding that this process will still require submission of paper documents until an electronic odometer solution is developed. Because continued submission of paper documents undermines the efficiencies sought, ―full‖ ERT – that is, all documents necessary for registration and titling should be capable of approval and/or acceptance by all parties, and should be capable of submission without transmittal or delivery of duplicate paper documents .– should remain the ultimate goal. ERT is not recommended as a means to eliminate review and approval of registration and titling transactions by the county treasurers, or to place registration and titling approval in the hands of the dealers, as county treasurers perform an important role in deterring fraud and promoting accuracy by determining the genuineness and regularity of each application. Authorizing dealers to act as registration agents that approve registration and title applications, issue registration receipts, and maintain and deliver permanent metal license plates is not recommended. Although distribution of permanent plates by dealers is not recommended, it is recommended that dealers participating in ERT generate and print registration applied for cards electronically. Unlike the manually-issued cards currently in use, cards issued in this fashion may be queried by law enforcement and are less susceptible to misuse by customers and dealers. The estimated time to implement the electronic application and registration applied for cards is 12 to 18 months, to begin after ELT and EFT have been implemented. It is recommended that focus during this time be on facilitating transfers through motor vehicle dealers, with initial deployment focused on higher-volume dealers that use DMS systems. In the long term an internet option for access to ERT must also be developed and maintained to allow participation for lower-volume dealers that do not use a DMS system. This option will also lay the ground work for an ERT option for sales between private individuals. Mandatory participation in Iowa is not recommended initially. As with ELT, it is recommended that mandatory participation be considered after at least an initial phase of ERT has been implemented and a suitable number of dealers have enrolled to provide a fair assessment of participation rates and opportunities. The use of vendors to facilitate ERT is not initially proposed because 1) DOT IT support staff is capable of developing a system that will interact with DMS systems and will still have to develop a dealer and public interface regardless of whether a vendor acts as intermediary between the DMS systems, and 2) there is concern that the cost of the vendor-based system, which is funded by transaction-based payments from the dealer to the vendor, will be passed to the consumer in the form of additional documentation or conveyance fees. However, the DOT recommends flexibility on this point, as development and pilot of the system may indicate that a multi-vendor approach similar to that recommended for ELT may increase the adoption rate by larger dealers and may ultimately decrease the user management to be exercised by DOT staff. If vendors are used in the process, additional legislation or administrative rules may be needed to control the fees that may be passed to the consumer. No direct cost to the DOT or county treasurers is expected, as the DOT expects that it may complete necessary programming with existing staff. Use of vendors to facilitate ERT transactions by dealers using DMS systems would result in transaction fees that may ultimately be passed to consumers. LEGISLATION – As a result of the changes implemented in 2004 under Senate File 2070, the only changes to Iowa statutes proposed are to section 321.69 of the Iowa Code, ―Damage disclosure statement,and section 321.71, ―Odometer requirements.‖ In each instance, authority to execute these statements by electronic means would be clarified by authorizing language similar to that used in section 321.20, subsections ―2‖ and ―3,‖ which allows for electronic applications and directs the department to ―adopt rules on the method for providing signatures for applications made by electronic means.‖ In these sections, the authorizing language might read as follows: Notwithstanding contrary provisions of this section, the department may develop and implement a program to allow for any statement required by this section to be made electronically. The department shall adopt rules on the method for providing signatures for statements made by electronic means. Some changes to DOT administrative rules will be useful but only to enable changes to work processes that would be desirable in the long term. Examples of long term work processes that would be enabled by rule changes include allowing for signatures created through electronic means and electronic odometer certifications. The DOT rules, as currently written, do not hinder the ability to proceed with ELT, EFT, and ERT.

Bi-planar 2D-to-3D registration in Fourier domain for stereoscopic X-ray motion tracking

In this paper we present a new method to track bonemovements in stereoscopic X-ray image series of the kneejoint. The method is based on two different X-ray imagesets: a rotational series of acquisitions of the stillsubject knee that will allow the tomographicreconstruction of the three-dimensional volume (model),and a stereoscopic image series of orthogonal projectionsas the subject performs movements. Tracking the movementsof bones throughout the stereoscopic image series meansto determine, for each frame, the best pose of everymoving element (bone) previously identified in the 3Dreconstructed model. The quality of a pose is reflectedin the similarity between its simulated projections andthe actual radiographs. We use direct Fourierreconstruction to approximate the three-dimensionalvolume of the knee joint. Then, to avoid the expensivecomputation of digitally rendered radiographs (DRR) forpose recovery, we reformulate the tracking problem in theFourier domain. Under the hypothesis of parallel X-raybeams, we use the central-slice-projection theorem toreplace the heavy 2D-to-3D registration of projections inthe signal domain by efficient slice-to-volumeregistration in the Fourier domain. Focusing onrotational movements, the translation-relevant phaseinformation can be discarded and we only consider scalarFourier amplitudes. The core of our motion trackingalgorithm can be implemented as a classical frame-wiseslice-to-volume registration task. Preliminary results onboth synthetic and real images confirm the validity ofour approach.

Fast Geodesic Active Fields for Image Registration Based on Splitting and Augmented Lagrangian Approaches.

In this paper, we present an efficient numerical scheme for the recently introduced geodesic active fields (GAF) framework for geometric image registration. This framework considers the registration task as a weighted minimal surface problem. Hence, the data-term and the regularization-term are combined through multiplication in a single, parametrization invariant and geometric cost functional. The multiplicative coupling provides an intrinsic, spatially varying and data-dependent tuning of the regularization strength, and the parametrization invariance allows working with images of nonflat geometry, generally defined on any smoothly parametrizable manifold. The resulting energy-minimizing flow, however, has poor numerical properties. Here, we provide an efficient numerical scheme that uses a splitting approach; data and regularity terms are optimized over two distinct deformation fields that are constrained to be equal via an augmented Lagrangian approach. Our approach is more flexible than standard Gaussian regularization, since one can interpolate freely between isotropic Gaussian and anisotropic TV-like smoothing. In this paper, we compare the geodesic active fields method with the popular Demons method and three more recent state-of-the-art algorithms: NL-optical flow, MRF image registration, and landmark-enhanced large displacement optical flow. Thus, we can show the advantages of the proposed FastGAF method. It compares favorably against Demons, both in terms of registration speed and quality. Over the range of example applications, it also consistently produces results not far from more dedicated state-of-the-art methods, illustrating the flexibility of the proposed framework.

Segmentation of Head and Neck Lymph Node Regions for Radiotherapy Planning Using Active Contour-Based Atlas Registration

In this paper, we present the segmentation of the headand neck lymph node regions using a new active contourbased atlas registration model. We propose to segment thelymph node regions without directly including them in theatlas registration process; instead, they are segmentedusing the dense deformation field computed from theregistration of the atlas structures with distinctboundaries. This approach results in robust and accuratesegmentation of the lymph node regions even in thepresence of significant anatomical variations between theatlas-image and the patient's image to be segmented. Wealso present a quantitative evaluation of lymph noderegions segmentation using various statistical as well asgeometrical metrics: sensitivity, specificity, dicesimilarity coefficient and Hausdorff distance. Acomparison of the proposed method with two other state ofthe art methods is presented. The robustness of theproposed method to the atlas selection, in segmenting thelymph node regions, is also evaluated.

Image Segmentation Using Information Bottleneck Method

In image processing, segmentation algorithms constitute one of the main focuses of research. In this paper, new image segmentation algorithms based on a hard version of the information bottleneck method are presented. The objective of this method is to extract a compact representation of a variable, considered the input, with minimal loss of mutual information with respect to another variable, considered the output. First, we introduce a split-and-merge algorithm based on the definition of an information channel between a set of regions (input) of the image and the intensity histogram bins (output). From this channel, the maximization of the mutual information gain is used to optimize the image partitioning. Then, the merging process of the regions obtained in the previous phase is carried out by minimizing the loss of mutual information. From the inversion of the above channel, we also present a new histogram clustering algorithm based on the minimization of the mutual information loss, where now the input variable represents the histogram bins and the output is given by the set of regions obtained from the above split-and-merge algorithm. Finally, we introduce two new clustering algorithms which show how the information bottleneck method can be applied to the registration channel obtained when two multimodal images are correctly aligned. Different experiments on 2-D and 3-D images show the behavior of the proposed algorithms

Fourier-based registration for robust forward-looking sonar mosaicing in low-visibility underwater environments

Vehicle operations in underwater environments are often compromised by poor visibility conditions. For instance, the perception range of optical devices is heavily constrained in turbid waters, thus complicating navigation and mapping tasks in environments such as harbors, bays, or rivers. A new generation of high-definition forward-looking sonars providing acoustic imagery at high frame rates has recently emerged as a promising alternative for working under these challenging conditions. However, the characteristics of the sonar data introduce difficulties in image registration, a key step in mosaicing and motion estimation applications. In this work, we propose the use of a Fourier-based registration technique capable of handling the low resolution, noise, and artifacts associated with sonar image formation. When compared to a state-of-the art region-based technique, our approach shows superior performance in the alignment of both consecutive and nonconsecutive views as well as higher robustness in featureless environments. The method is used to compute pose constraints between sonar frames that, integrated inside a global alignment framework, enable the rendering of consistent acoustic mosaics with high detail and increased resolution. An extensive experimental section is reported showing results in relevant field applications, such as ship hull inspection and harbor mapping

Multimodal image registration of the scoliotic torso for surgical planning

Background This paper presents a method that registers MRIs acquired in prone position, with surface topography (TP) and X-ray reconstructions acquired in standing position, in order to obtain a 3D representation of a human torso incorporating the external surface, bone structures, and soft tissues. Methods TP and X-ray data are registered using landmarks. Bone structures are used to register each MRI slice using an articulated model, and the soft tissue is confined to the volume delimited by the trunk and bone surfaces using a constrained thin-plate spline. Results The method is tested on 3 pre-surgical patients with scoliosis and shows a significant improvement, qualitatively and using the Dice similarity coefficient, in fitting the MRI into the standing patient model when compared to rigid and articulated model registration. The determinant of the Jacobian of the registration deformation shows higher variations in the deformation in areas closer to the surface of the torso. Conclusions The novel, resulting 3D full torso model can provide a more complete representation of patient geometry to be incorporated in surgical simulators under development that aim at predicting the effect of scoliosis surgery on the external appearance of the patient’s torso.

Articulated Model Registration of MRI/X-Ray Spine Data

This paper presents a method based on articulated models for the registration of spine data extracted from multimodal medical images of patients with scoliosis. With the ultimate aim being the development of a complete geometrical model of the torso of a scoliotic patient, this work presents a method for the registration of vertebral column data using 3D magnetic resonance images (MRI) acquired in prone position and X-ray data acquired in standing position for five patients with scoliosis. The 3D shape of the vertebrae is estimated from both image modalities for each patient, and an articulated model is used in order to calculate intervertebral transformations required in order to align the vertebrae between both postures. Euclidean distances between anatomical landmarks are calculated in order to assess multimodal registration error. Results show a decrease in the Euclidean distance using the proposed method compared to rigid registration and more physically realistic vertebrae deformations compared to thin-plate-spline (TPS) registration thus improving alignment.

2D-3D Rigid-Body Registration of X-Ray Fluoroscopy and CT Images

The registration of pre-operative volumetric datasets to intra- operative two-dimensional images provides an improved way of verifying patient position and medical instrument loca- tion. In applications from orthopedics to neurosurgery, it has a great value in maintaining up-to-date information about changes due to intervention. We propose a mutual information- based registration algorithm to establish the proper align- ment. For optimization purposes, we compare the perfor- mance of the non-gradient Powell method and two slightly di erent versions of a stochastic gradient ascent strategy: one using a sparsely sampled histogramming approach and the other Parzen windowing to carry out probability density approximation. Our main contribution lies in adopting the stochastic ap- proximation scheme successfully applied in 3D-3D registra- tion problems to the 2D-3D scenario, which obviates the need for the generation of full DRRs at each iteration of pose op- timization. This facilitates a considerable savings in compu- tation expense. We also introduce a new probability density estimator for image intensities via sparse histogramming, de- rive gradient estimates for the density measures required by the maximization procedure and introduce the framework for a multiresolution strategy to the problem. Registration results are presented on uoroscopy and CT datasets of a plastic pelvis and a real skull, and on a high-resolution CT- derived simulated dataset of a real skull, a plastic skull, a plastic pelvis and a plastic lumbar spine segment.

Compact Representations for Fast Nonrigid Registration of Medical Images

We develop efficient techniques for the non-rigid registration of medical images by using representations that adapt to the anatomy found in such images. Images of anatomical structures typically have uniform intensity interiors and smooth boundaries. We create methods to represent such regions compactly using tetrahedra. Unlike voxel-based representations, tetrahedra can accurately describe the expected smooth surfaces of medical objects. Furthermore, the interior of such objects can be represented using a small number of tetrahedra. Rather than describing a medical object using tens of thousands of voxels, our representations generally contain only a few thousand elements. Tetrahedra facilitate the creation of efficient non-rigid registration algorithms based on finite element methods (FEM). We create a fast, FEM-based method to non-rigidly register segmented anatomical structures from two subjects. Using our compact tetrahedral representations, this method generally requires less than one minute of processing time on a desktop PC. We also create a novel method for the non-rigid registration of gray scale images. To facilitate a fast method, we create a tetrahedral representation of a displacement field that automatically adapts to both the anatomy in an image and to the displacement field. The resulting algorithm has a computational cost that is dominated by the number of nodes in the mesh (about 10,000), rather than the number of voxels in an image (nearly 10,000,000). For many non-rigid registration problems, we can find a transformation from one image to another in five minutes. This speed is important as it allows use of the algorithm during surgery. We apply our algorithms to find correlations between the shape of anatomical structures and the presence of schizophrenia. We show that a study based on our representations outperforms studies based on other representations. We also use the results of our non-rigid registration algorithm as the basis of a segmentation algorithm. That algorithm also outperforms other methods in our tests, producing smoother segmentations and more accurately reproducing manual segmentations.

Medical image registration based on random line sampling

One of the key aspects in 3D-image registration is the computation of the joint intensity histogram. We propose a new approach to compute this histogram using uniformly distributed random lines to sample stochastically the overlapping volume between two 3D-images. The intensity values are captured from the lines at evenly spaced positions, taking an initial random offset different for each line. This method provides us with an accurate, robust and fast mutual information-based registration. The interpolation effects are drastically reduced, due to the stochastic nature of the line generation, and the alignment process is also accelerated. The results obtained show a better performance of the introduced method than the classic computation of the joint histogram