Bootstrap resampling for image registration uncertainty estimation without ground truth

We address the problem of estimating the uncertainty of pixel based image registration algorithms, given just the two images to be registered, for cases when no ground truth data is available. Our novel method uses bootstrap resampling. It is very general, applicable to almost any registration method based on minimizing a pixel-based similarity criterion; we demonstrate it using the SSD, SAD, correlation, and mutual information criteria. We show experimentally that the bootstrap method provides better estimates of the registration accuracy than the state-of-the-art CramEr-Rao bound method. Additionally, we evaluate also a fast registration accuracy estimation (FRAE) method which is based on quadratic sensitivity analysis ideas and has a negligible computational overhead. FRAE mostly works better than the CramEr-Rao bound method but is outperformed by the bootstrap method.

Symmetric and transitive registration of image sequences

This paper presents a method for constructing symmetric and transitive algorithms for registration of image sequences from image registration algorithms that do not have these two properties. The method is applicable to both rigid and nonrigid registration and it can be used with linear or periodic image sequences. The symmetry and transitivity properties are satisfied exactly (up to the machine precision), that is, they always hold regardless of the image type, quality, and the registration algorithm as long as the computed transformations are invertable. These two properties are especially important in motion tracking applications since physically incorrect deformations might be obtained if the registration algorithm is not symmetric and transitive. The method was tested on two sequences of cardiac magnetic resonance images using two different nonrigid image registration algorithms. It was demonstrated that the transitivity and symmetry errors of the symmetric and transitive modification of the algorithms could be made arbitrary small when the computed transformations are invertable, whereas the corresponding errors for the nonmodified algorithms were on the order of the pixel size. Furthermore, the symmetric and transitive modification of the algorithms had higher registration accuracy than the nonmodified algorithms for both image sequences.

Automated identification of neurons in 3D confocal datasets from zebrafish brainstem

Many kinds of neuroscience data are being acquired regarding the dynamic behaviour and phenotypic diversity of nerve cells. But as the size, complexity and numbers of 3D neuroanatomical datasets grow ever larger, the need for automated detection and analysis of individual neurons takes on greater importance. We describe here a method that detects and identifies neurons within confocal image stacks acquired from the zebrafish brainstem. The first step is to create a template that incorporates the location of all known neurons within a population - in this case the population of reticulospinal cells. Once created, the template is used in conjunction with a sequence of algorithms to determine the 3D location and identity of all fluorescent neurons in each confocal dataset. After an image registration step, neurons are segmented within the confocal image stack and subsequently localized to specific locations within the brainstem template - in many instances identifying neurons as specific, individual reticulospinal cells. This image-processing sequence is fully automated except for the initial selection of three registration points on a maximum projection image. In analysing confocal image stacks that ranged considerably in image quality, we found that this method correctly identified on average approximately 80% of the neurons (if we assume that manual detection by experts constitutes 'ground truth'). Because this identification can be generated approximately 100 times faster than manual identification, it offers a considerable time savings for the investigation of zebrafish reticulospinal neurons. In addition to its cell identification function, this protocol might also be integrated with stereological techniques to enhance quantification of neurons in larger databases. Our focus has been on zebrafish brainstem systems, but the methods described should be applicable to diverse neural architectures including retina, hippocampus and cerebral cortex.

Automatic registration of serial sections of mouse lymph node by using Image-Reg

The first step towards the three-dimensional (3D) reconstruction of histological structures from serial sectioned tissue blocks is the proper alignment of microscope image sequences. We have accomplished an automatic rigid registration program, named Image-Reg, to align serial sections from mouse lymph node and Peyer's patch. Our approach is based on the calculation of the pixel-correlation of objects in adjacent images. The registration process is mainly divided into two steps. Once the foreground images have been segmented from the original images, the first step (primary alignment) is performed on the binary images of segmented objects; this process includes rotation by using the moments and translation through the X, Y axes by using the centroid. In the second step, the matching error of two binary images is calculated and the registration results are refined through multi-scale iterations. In order to test the registration performance, Image-Reg has been applied to an image and its transformed (rotated) version and subsequently to an image sequence of three serial sections of mouse lymph node. In addition, to compare our algorithm with other registration methods, three other approaches, viz. manual registration with Reconstruct, semi-automatic landmark registration with Image-Pro Plus and the automatic phase-correlation method with Image-Pro Plus, have also been applied to these three sections. The performance of our program has been also tested on other two-image data sets. These include: (a) two light microscopic images acquired by the automatic microscope (stitched with other software); (b) two images fluorescent images acquired by confocal microscopy (tiled with other software). Our proposed approach provides a fast and accurate linear alignment of serial image sequences for the 3D reconstruction of tissues and organs.

Evolving role of imaging modalities in inflammatory breast cancer

Imaging plays several key roles in the diagnosis and assessment of inflammatory breast cancer (IBC). These include characterization of the known tumor, delineation of locoregional disease in the ipsilateral and contralateral breast and regional lymph node basins, diagnosis of distant metastases, and evaluation of treatment response. We review the role of conventional imaging modalities, including mammography and sonography. We also discuss the potential of using evolving imaging modalities such as magnetic resonance imaging (MRI), positron emission tomography with computed tomography (PET/CT), and more advanced or emerging imaging techniques in the assessment of IBC.

Towards a noninvasive intracranial tumor irradiation using 3d optical imaging and multimodal data registration

Conformal radiotherapy (CRT) results in high-precision tumor volume irradiation. In fractioned radiotherapy (FRT), lesions are irradiated in several sessions so that healthy neighbouring tissues are better preserved than when treatment is carried out in one fraction. In the case of intracranial tumors, classical methods of patient positioning in the irradiation machine coordinate system are invasive and only allow for CRT in one irradiation session. This contribution presents a noninvasive positioning method representing a first step towards the combination of CRT and FRT. The 3D data used for the positioning is point clouds spread over the patient's head (CT-data usually acquired during treatment) and points distributed over the patient's face which are acquired with a structured light sensor fixed in the therapy room. The geometrical transformation linking the coordinate systems of the diagnosis device (CT-modality) and the 3D sensor of the therapy room (visible light modality) is obtained by registering the surfaces represented by the two 3D point sets. The geometrical relationship between the coordinate systems of the 3D sensor and the irradiation machine is given by a calibration of the sensor position in the therapy room. The global transformation, computed with the two previous transformations, is sufficient to predict the tumor position in the irradiation machine coordinate system with only the corresponding position in the CT-coordinate system. Results obtained for a phantom show that the mean positioning error of tumors on the treatment machine isocentre is 0.4 mm. Tests performed with human data proved that the registration algorithm is accurate (0.1 mm mean distance between homologous points) and robust even for facial expression changes.

Cardiac image integration implications for atrial fibrillation ablation

Cardiac image registration using computed tomography (CT), magnetic resonance imaging (MRI), ultrasound, and fluoroscopy is currently being investigated and clinically used for atrial fibrillation (AF) ablation. Cardiac image registration, in the context of left atrium, is intermodal, with the acquired image and the real-time reference image residing in different image spaces, and involves optimization, where one image space is transformed into the other. Geometry-based methods, which include using fiducial points and/or surface-based techniques, are usually used for cardiac image registration. During fiducial point registration, fiducial points are either external skin markers or manually determined by marking anatomical landmarks, using mapping catheters. Usually, a minimum of three non collinear points are needed for optimal registration. Recently, a catheter placed inside the coronary sinus has also been used as a fiducial marker for the purpose of registration. During surface registration, the process involves characterizing surfaces in each of the images and deriving the best transformation between them. Unlike rigid body registration, such as has been extensively used in imaging the brain, cardiac image registration is unique and challenging. In addition to the errors inherent in intermodal registration, such as errors in pixel and voxel dimension and errors due to fiducial point selection, there are errors specific to cardiac image registration, i.e., errors due to cardiac motion during the cardiac cycle and due to respiration. This review addresses the basic principles of registration and the inherent registration errors as they relate to cardiac imaging and registration.

Field-warp registration for biomedical high-resolution thermal infrared images

Biomedical protocols based on thermal infrared images often require effective image registration. Algorithms specifically designed for registration of thermal images are hardly available and use of algorithms designed for other imaging techniques may result poorly reliable. In fact, registration algorithms developed for other biomedical images often require rigid-body assumption or limited range for intensity values. Such assumption may not be applicable for human body thermal images. Therefore, we present here an adaptation of a field-warp based method as a possible solution for thermal image registration. The method was applied for registering images taken from an experimental protocol aimed at comparing total body skin temperature distribution in natural or altered posture. The method appears to be effective into providing a reliable tool for objective intra and inter individual skin temperature distribution comparisons.

Targeting liver lesions for radiofrequency ablation: an experimental feasibility study using a CT-US fusion imaging system

PURPOSE

To investigate the feasibility and validity of real-time guidance using a fusion imaging system that combines ultrasound (US) and computed tomography (CT) in the targeting and subsequent radiofrequency (RF) ablation of a liver target inconspicuous on US.

METHODS AND MATERIALS

The study was designed as an experimental ex vivo study in calf livers with radiopaque internal targets, inconspicuous at US, simulating a focal liver lesion. The study included 2 phases. The initial phase was to examine the feasibility of matching preprocedural volumetric CT data of the calf livers with real-time US using a commercially available multimodality fusion imaging system (Virtual Navigator System, Esaote SpA, Genoa, Italy), and to assess the accuracy of targeting using a 22 gauge cytologic needle. The second phase of the study was to validate such a technique using a 15 gauge RF multitined expandable needle (RITA Medical Systems, Mountain View, CA) and to examine the accuracy of the needle placement relative to the target. The tip of the trocar of the RF needle had to be placed 1 cm from the target and then the hooks had to be deployed to 3 cm. Unenhanced CT of the liver and multiplanar reconstructions were performed to calculate accuracy of positioning, ie, the lateral distance between the needle and the target, the distance between the tip of the trocar of the RF electrode and the target, and the lateral distance between the central tine of the RF electrode and the target.

RESULTS

All calf livers underwent successful CT-US registration with a mean registration error of 3.0 +/- 0.1 mm and 2.9 +/- 0.1 mm in the initial and second phase of the study, respectively. In the initial phase an overall number of 24 insertions were performed after the US-CT guidance. The mean needle to target distance was 1.9 +/- 0.7 mm (range, 0.8-3.0 mm). In the second phase an overall number of 12 ablations were performed. The mean target-trocar distance was 10.3 +/- 2.6 mm. The mean target-central tine lateral distance was 3.9 +/- 0.7 mm (range, 2.9-5.1 mm). After the dissection of the specimen the target was found unchanged in the center of the ablation zone in all cases.

CONCLUSION

Real-time registration and fusion of preprocedure CT volume images with intraprocedure US is feasible and accurate. The study was however conducted in an ideal experimental setting, without patient movements and breathing, and further studies are warranted to validate the system under clinical conditions.

Strain fields in histological slices of brain tissue determined by synchrotron radiation-based micro computed tomography

Accurate knowledge of the morphology of the human brain is required for minimally or non-invasive surgical interventions. On the (sub-)cellular level, brain tissue is generally characterized using optical microscopy, which allows extracting morphological features with a wide spectrum of staining procedures. The preparation of the histological slices, however, often leads to artifacts resulting in imperfect morphological data. In addition, the generation of 3D data is time-consuming. Therefore, we propose synchrotron radiation-based micro computed tomography (SRmicroCT) avoiding preparation artifacts and giving rise to the 3D morphology of features such as gray and white matter on the micrometer level. One can differentiate between white and gray matter without any staining procedure because of different X-ray absorption values. At the photon energy of 10keV, the white matter exhibits the absorption of 5.08 cm(-1), whereby the value for the gray matter corresponds to 5.25 cm(-1). The tomography data allow quantifying the local strains in the histological images using registration algorithms. The deformation of histological slices compared to the SRmicroCT in a 2D-2D registration leads to values of up to 6.3%. Mean deformation values for the Nissl-stained slices are determined to about 1%, whereas the myelin-stained slices yield slightly higher values than 2%.

A review of geometric transformations for nonrigid body registration

This paper provides a comprehensive and quantitative review of spatial transformations models for nonrigid image registration. It explains the theoretical foundation of the models and classifies them according to this basis. This results in two categories, physically based models described by partial differential equations of continuum mechanics (e.g., linear elasticity and fluid flow) and basis function expansions derived from interpolation and approximation theory (e.g., radial basis functions, B-splines and wavelets). Recent work on constraining the transformation so that it preserves the topology or is diffeomorphic is also described. The final section reviews some recent evaluation studies. The paper concludes by explaining under what conditions a particular transformation model is appropriate.

Multiscale joint segmentation and registration of image morphology

Multimodal image registration significantly benefits from previous denoising and structure segmentation and vice versa. In particular combined information of different image modalities makes segmentation significantly more robust. Indeed, fundamental tasks in image processing are highly interdependent. A variational approach is presented, which combines the detection of corresponding edges, an edge preserving denoising and the morphological registration via a non-rigid deformation for a pair of images with structural correspondence. The morphology of an image function is split into a singular part consisting of the edge set and a regular part represented by the field of normals on the ensemble of level sets. A Mumford-Shah type free discontinuity problem is applied to treat the singular morphology and the matching of corresponding edges under the deformation. The matching of the regular morphology is quantified by a second contribution which compares deformed normals and normals at deformed positions. Finally, a nonlinear elastic energy controls the deformation itself and ensures smoothness and injectivity. A multi scale approach that is based on a phase field approximation leads to an effective and efficient algorithm. Numerical experiments underline the robustness of the presented approach and show applications on medical images.

Image-to-patient registration techniques in head surgery

Frame-based stereotaxy was developed in neurosurgery at the beginning of the last century, evolving from atlas-based stereotaxy to stereotaxy based on the individual patient's image data. This established method is still in use in neurosurgery and radiotherapy. There have since been two main developments based on this concept: frameless stereotaxy and markerless registration. Frameless stereotactic systems ('navigation systems') replaced the cumbersome stereotactic frame by mechanically and later also optically or magnetically tracked instruments. Stereotaxy based on the individual patient's image data introduced the problem of patient-to-image data registration. The development of navigation systems based on frameless stereotaxy has dramatically increased its use in surgical disciplines other than neurosurgery, but image-guided surgery based on fiducial marker registration needs dedicated imaging for registration purposes, in addition to the diagnostic imaging that might have been performed. Markerless registration techniques can overcome the resulting additional cost and effort, and result in more widespread use of image-guided surgery techniques. In this review paper, the developments that led to today's navigation systems are outlined, and the applications and possibilities of these methods in the field of maxillofacial surgery are presented.

A method of image registration based on its geometric character

This paper aimed at building a model for the image registration of CT image and slice image. We used a registration method based on the geometric character of the outline of the image. This method need us do pretreatment to the image, in order to get rid of the noise and enhance the outline of the image, here we use time-frequency theory such as wavelet transformation a gray image fusion etc. The result of image registration shows that this method is fast a simple and effective.

Image registration for breast imaging: a review

Breast cancer is the most common type of cancer in women worldwide. About ten percent of women are confronted with breast cancer in their lives. Breast Cancer can be most efficiently treated if detected at an early sage. Imaging of the breast can be accomplished using several modalities such as: X-ray, MRI, CT, Ultrasound, and now Molecular Imaging. Image registration plays a critical role in breast imaging. It provides aid to better visualization of lesions on bilateral or temporal X-ray mammograms, or in the fusion of different modalities acquired using different principles of physics. The non-rigid, inhomogeneous, anisotropic and temporally changing nature of breast tissue make breast image registration a challenging task. This paper presents an overview of the current state-of-the-art in the breast image registration techniques. Methods are classified according to the modalities involved in the registration process. Intra-modality registration techniques focus on X-ray mammogram registration, while inter-modality techniques will cover the registration of X-ray with other modality. Validation of breast registration methods is also discussed.

Image Registration in Intra-oral Radiography

Image registration is one of the image processing methods which is widely used in computer vision, pattern recognition, and medical imaging. In digital subtraction radiography, image registration is one of the important prerequisites to match the reference and subsequent images. In this paper, we propose an automatic non-rigid registration method namely curvature-based registration that relies on a curvature based penalizing term and its application on dental radiography. The regularizing term of this intensity-based registration approach provides affine linear transformation so that pre-registration step is no longer necessary. This leads to faster and more reliable solutions. The implementation of this approach is based on the numerical solution of the underlying Euler-Lagrange equations. In addition, a comparison between this algorithm and Linear Alignment Method (LAM) with 20 image pairs is presented.

Validation of an elastic matching algorithm based on complex wavelets for the realignment of dynamic MR breast images

In this paper, a novel approach for the registration of dynamic magnetic resonance (MR) breast images is validated. The approach is fully automatic and it performs a multi-resolution motion field estimation based on complex discrete wavelets transform (CDWT). The validation was designed to assess the registration quality in patient breast data. A set of diagnostic features of MR breast image was defined and two expert radiologists were asked to score the capability to detect these features in the subtraction images. Two registration methods were compared: a rigid registration algorithm and the proposed approach. Results show that the novel approach is superior in terms of both lesion detectability and lesion definition.

Implementation and evaluation of a new workflow for registration and segmentation of pulmonary MRI data for regional lung perfusion assessment

Recently it has been shown that regional lung perfusion can be assessed using time-resolved contrast-enhanced magnetic resonance (MR) imaging. Quantification of the perfusion images has been attempted, based on definition of small regions of interest (ROIs). Use of complete lung segmentations instead of ROIs could possibly increase quantification accuracy. Due to the low signal-to-noise ratio, automatic segmentation algorithms cannot be applied. On the other hand, manual segmentation of the lung tissue is very time consuming and can become inaccurate, as the borders of the lung to adjacent tissues are not always clearly visible. We propose a new workflow for semi-automatic segmentation of the lung from additionally acquired morphological HASTE MR images. First the lung is delineated semi-automatically in the HASTE image. Next the HASTE image is automatically registered with the perfusion images. Finally, the transformation resulting from the registration is used to align the lung segmentation from the morphological dataset with the perfusion images. We evaluated rigid, affine and locally elastic transformations, suitable optimizers and different implementations of mutual information (MI) metrics to determine the best possible registration algorithm. We located the shortcomings of the registration procedure and under which conditions automatic registration will succeed or fail. Segmentation results were evaluated using overlap and distance measures. Integration of the new workflow reduces the time needed for post-processing of the data, simplifies the perfusion quantification and reduces interobserver variability in the segmentation process. In addition, the matched morphological data set can be used to identify morphologic changes as the source for the perfusion abnormalities.

Single-Photon Emission Computed Tomography/Computed Tomography in Brain Tumors

Anatomic imaging procedures (computed tomography [CT] and magnetic resonance imaging [MRI]) have become essential tools for brain tumor assessment. Functional images (positron emission tomography [PET] and single-photon emission computed tomography [SPECT]) can provide additional information useful during the diagnostic workup to determine the degree of malignancy and as a substitute or guide for biopsy. After surgery and/or radiotherapy, nuclear medicine examinations are essential to assess persistence of tumor, to differentiate recurrence from radiation necrosis and gliosis, and to monitor the disease. The combination of functional images with anatomic ones is of the utmost importance for a full evaluation of these patients, which can be obtained by means of imaging fusion. Despite the fast-growing diffusion of PET, in most cases of brain tumors, SPECT studies are adequate and provide results that parallel those obtained with PET. The main limitation of SPECT imaging with brain tumor-seeking radiopharmaceuticals is the lack of precise anatomic details; this drawback is overcome by the fusion with morphological studies that provide an anatomic map to scintigraphic data. In the past, software-based fusion of independently performed SPECT and CT or MRI demonstrated usefulness for brain tumor assessment, but this process is often time consuming and not practical for everyday nuclear medicine studies. The recent development of dual-modality integrated imaging systems, which allow the acquisition of SPECT and CT images in the same scanning session, and their co-registration by means of the hardware, has facilitated this process. In SPECT studies of brain tumors with various radiopharmaceuticals, fused images are helpful in providing the precise localization of neoplastic lesions, and in excluding the disease in sites of physiologic tracer uptake. This information is useful for optimizing diagnosis, therapy monitoring, and radiotherapy treatment planning, with a positive impact on patient management.

Statistical neuroanatomy of the human inferior frontal gyrus and probabilistic atlas in a standard stereotaxic space

We manually defined the inferior frontal gyrus (IFG) on high-resolution MRIs in native space in 30 healthy subjects (15 female, median age 31 years; 15 male, median age 30 years), resulting in 30 individual atlases. Using standard software (SPM99), these were spatially transformed to a widely used stereotaxic space (MNI/ICBM 152) to create probabilistic maps. In native space, the total IFG volume was on average 5%, and the gray matter (GM) portion 12% larger in women (not significant). Expressed as a percentage of ipsilateral frontal lobe volume (i.e., correcting for brain size), the IFG was an average of 20%, and the GM portion of the IFG 27%, larger in women (P < 0.005). Correcting for total lobar volume yielded the same result. No asymmetry was found in IFG volumes. There were significant positional differences between the right and left IFGs, with the right IFG being further lateral in both native and stereotaxic space. Variability was similar on the left and right, but more pronounced anteriorly and superiorly. We show differences in IFG volume, composition, and position between sexes and between hemispheres. Applications include probabilistic determination of location in group studies, automatic labeling of new scans, and detection of anatomical abnormalities in patients.

Registration of 3D computed tomographic images with interventional systems: Implications for catheter ablation of atrial fibrillation

Despite the great promise catheter ablation offers in the treatment of complex arrhythmias such as atrial fibrillation (AF), long procedure times and somewhat suboptimal results hinder the widespread use of this technique. As fluoroscopy does not provide contrast differentiation between the area of interest and the surrounding structures, there is a lack of proper intra procedure image guidance. Segmentation of anatomical structures such as the left atrium (LA) can be performed using images obtained with modalities such as computed tomography (CT). However, unlike the cardiac mapping systems, these imaging systems do not track catheters in real time. This review addresses the evolving concept of image registration to deliver therapy in cardiac arrhythmias.

A cross validation study of deep brain stimulation targeting: from experts to atlas-based, segmentation-based and automatic registration algorithms

Validation of image registration algorithms is a difficult task and open-ended problem, usually application-dependent. In this paper, we focus on deep brain stimulation (DBS) targeting for the treatment of movement disorders like Parkinson's disease and essential tremor. DBS involves implantation of an electrode deep inside the brain to electrically stimulate specific areas shutting down the disease's symptoms. The subthalamic nucleus (STN) has turned out to be the optimal target for this kind of surgery. Unfortunately, the STN is in general not clearly distinguishable in common medical imaging modalities. Usual techniques to infer its location are the use of anatomical atlases and visible surrounding landmarks. Surgeons have to adjust the electrode intraoperatively using electrophysiological recordings and macrostimulation tests. We constructed a ground truth derived from specific patients whose STNs are clearly visible on magnetic resonance (MR) T2-weighted images. A patient is chosen as atlas both for the right and left sides. Then, by registering each patient with the atlas using different methods, several estimations of the STN location are obtained. Two studies are driven using our proposed validation scheme. First, a comparison between different atlas-based and nonrigid registration algorithms with a evaluation of their performance and usability to locate the STN automatically. Second, a study of which visible surrounding structures influence the STN location. The two studies are cross validated between them and against expert's variability. Using this scheme, we evaluated the expert's ability against the estimation error provided by the tested algorithms and we demonstrated that automatic STN targeting is possible and as accurate as the expert-driven techniques currently used. We also show which structures have to be taken into account to accurately estimate the STN location.

Subpixel estimation of shifts directly in the Fourier domain

In this paper, we establish the exact relationship between the continuous and the discrete phase difference of two shifted images, and show that their discrete phase difference is a two-dimensional sawtooth signal. Subpixel registration can, thus, be performed directly in the Fourier domain by counting number of cycles of the phase difference matrix along each frequency axis. The subpixel portion is given by the noninteger fraction of the last cycle along each axis. The problem is formulated as an overdetermined homogeneous quadratic cost function under rank constraint for the phase difference, and the shape constraint for the filter that computes the group delay. The optimal tradeoff for imposing the constraints is determined using the method of generalized cross validation. Also, in order to robustify the solution, we assume a mixture model of inlying and outlying estimated shifts and truncate our quadratic cost function using expectation maximization.

Medical image registration with partial data

We have developed a general-purpose registration algorithm for medical images and volumes. The transformation between images is modeled as locally affine but globally smooth, and explicitly accounts for local and global variations in image intensities. An explicit model of missing data is also incorporated, allowing us to simultaneously segment and register images with partial or missing data. The algorithm is built upon a differential multiscale framework and incorporates the expectation maximization algorithm. We show that this approach is highly effective in registering a range of synthetic and clinical medical images.

Fusion of radiostereometric analysis data into computed tomography space: application to the elbow joint

Improvement of joint prostheses is dependent upon information concerning the biomechanical properties of the joint. Radiostereometric analysis (RSA) and electromagnetic techniques have been applied in previous cadaver and in vivo studies on the elbow joint to provide valuable information concerning joint motion axes. However, such information is limited to mathematically calculated positions of the axes according to an orthogonal coordinate system and is difficult to relate to individual skeletal anatomy. The aim of this study was to evaluate the in vivo application of a new fusion method to provide three-dimensional (3D) visualization of flexion axes according to bony landmarks. In vivo RSA data of the elbow joint's flexion axes was combined with data obtained by 3D computed tomography (CT). Results were obtained from five healthy subjects after one was excluded due to an instable RSA marker. The median error between imported and transformed RSA marker coordinates and those obtained in the CT volume was 0.22 mm. Median maximal rotation error after transformation of the rigid RSA body to the CT volume was 0.003 degrees . Points of interception with a plane calculated in the RSA orthogonal coordinate system were imported into the CT volume, facilitating the 3D visualization of the flexion axes. This study demonstrates a successful fusion of RSA and CT data, without significant loss of RSA accuracy. The method could be used for relating individual motion axes to a 3D representation of relevant joint anatomy, thus providing important information for clinical applications such as the development of joint prostheses.

Hybrid retinal image registration

This work studies retinal image registration in the context of the National Institutes of Health (NIH) Early Treatment Diabetic Retinopathy Study (ETDRS) standard. The ETDRS imaging protocol specifies seven fields of each retina and presents three major challenges for the image registration task. First, small overlaps between adjacent fields lead to inadequate landmark points for feature-based methods. Second, the non-uniform contrast/intensity distributions due to imperfect data acquisition will deteriorate the performance of area-based techniques. Third, high-resolution images contain large homogeneous nonvascular/texureless regions that weaken the capabilities of both feature-based and area-based techniques. In this work, we propose a hybrid retinal image registration approach for ETDRS images that effectively combines both area-based and feature-based methods. Four major steps are involved. First, the vascular tree is extracted by using an efficient local entropy-based thresholding technique. Next, zeroth-order translation is estimated by maximizing mutual information based on the binary image pair (area-based). Then image quality assessment regarding the ETDRS field definition is performed based on the translation model. If the image pair is accepted, higher-order transformations will be involved. Specifically, we use two types of features, landmark points and sampling points, for affine/quadratic model estimation. Three empirical conditions are derived experimentally to control the algorithm progress, so that we can achieve the lowest registration error and the highest success rate. Simulation results on 504 pairs of ETDRS images show the effectiveness and robustness of the proposed algorithm.

Nonrigid medical image registration by finite-element deformable sheet-curve models

Image-based change quantitation has been recognized as a promising tool for accurate assessment of tumor's early response to chemoprevention in cancer research. For example, various changes on breast density and vascularity in glandular tissue are the indicators of early response to treatment. Accurate extraction of glandular tissue from pre- and postcontrast magnetic resonance (MR) images requires a nonrigid registration of sequential MR images embedded with local deformations. This paper reports a newly developed registration method that aligns MR breast images using finite-element deformable sheet-curve models. Specifically, deformable curves are constructed to match the boundaries dynamically, while a deformable sheet of thin-plate splines is designed to model complex local deformations. The experimental results on both digital phantoms and real MR breast images using the new method have been compared to point-based thin-plate-spline (TPS) approach, and have demonstrated a significant and robust improvement in both boundary alignment and local deformation recovery.

Feasibility and validation of registration of three-dimensional left atrial models derived from computed tomography with a noncontact cardiac mapping system

OBJECTIVES

The purpose of this study was to determine the feasibility and assess the validity of registering three-dimensional (3D) models from computed tomographic (CT) images using a cardiac mapping system.

BACKGROUND

Registration of 3D anatomic models with an interventional system could help identify and navigate mapping and ablation catheters over a complex structure such as the left atrium (LA).

METHODS

ECG-gated, contrast-enhanced cardiac CT imaging was performed in 14 patients with atrial fibrillation. Segmentation was used to create 3D models of the LA. The 3D models were registered with the mapping system using a series of fiducial points. Registration was accomplished retrospectively in the first 10 patients, and catheter navigation was visualized from recorded data. In the final four patients, registration was accomplished in real time during electrophysiologic study. The mapping catheter position, as it was navigated inside the LA, was applied to the registered model in real time. For the validation study, temporary pacing leads were implanted in the LA of 10 dogs. Following this, CT scanning, segmentation, LA model importation, and registration was described previously. After registration, a mapping catheter was positioned at the site of each buried lead according to the registered model with no fluoroscopic guidance. A radiofrequency lesion was created at this location, and the dog was sacrificed, the heart removed and stained, and the distance between the buried lead and the lesion measured.

RESULTS

During the feasibility study, the location of the catheter in the registered model correlated with fluoroscopy, angiography, and intracardiac electrograms. LA endocardial potentials during sinus rhythm and any premature atrial contractions also were successfully delineated over the registered models. In the validation study, the mean target registration error was 2.0 +/- 3.6 mm.

CONCLUSIONS

Registration of CT-derived 3D models of the LA using a cardiac mapping system is feasible and accurate.

Piecewise affine registration of biological images for volume reconstruction

This manuscript tackles the reconstruction of 3-D volumes via mono-modal registration of series of 2-D biological images (histological sections, autoradiographs, cryosections, etc.). The process of acquiring these images typically induces composite transformations that we model as a number of rigid or affine local transformations embedded in an elastic one. We propose a registration approach closely derived from this model. Given a pair of input images, we first compute a dense similarity field between them with a block matching algorithm. We use as a similarity measure an extension of the classical correlation coefficient that improves the consistency of the field. A hierarchical clustering algorithm then automatically partitions the field into a number of classes from which we extract independent pairs of sub-images. Our clustering algorithm relies on the Earth mover's distribution metric and is additionally guided by robust least-square estimation of the transformations associated with each cluster. Finally, the pairs of sub-images are, independently, affinely registered and a hybrid affine/non-linear interpolation scheme is used to compose the output registered image. We investigate the behavior of our approach on several batches of histological data and discuss its sensitivity to parameters and noise.

Reconstruct: a free editor for serial section microscopy

Many microscopy studies require reconstruction from serial sections, a method of analysis that is sometimes difficult and time-consuming. When each section is cut, mounted and imaged separately, section images must be montaged and realigned to accurately analyse and visualize the three-dimensional (3D) structure. Reconstruct is a free editor designed to facilitate montaging, alignment, analysis and visualization of serial sections. The methods used by Reconstruct for organizing, transforming and displaying data enable the analysis of series with large numbers of sections and images over a large range of magnifications by making efficient use of computer memory. Alignments can correct for some types of non-linear deformations, including cracks and folds, as often encountered in serial electron microscopy. A large number of different structures can be easily traced and placed together in a single 3D scene that can be animated or saved. As a flexible editor, Reconstruct can reduce the time and resources expended for serial section studies and allows a larger tissue volume to be analysed more quickly.