Search

The purpose of this study is to quantitatively compare the accuracy of spatial registration of Cartesian breath-hold 3D-GRE and non-respiratory-triggered free-breathing radial 3D-GRE images with PET data acquisition on whole-body hybrid MR–PET system. Eight patients (six men and two women; mean age, 56.6 ± 5.5 years) with nine ablated hepatocellular carcinomas constituted our study population. Spatial coordinates (x, y, z) of the estimated isocenters of the ablated areas were independently determined by two radiologists. Both T1-weighted sequences were performed in the axial plane. Distance between the isocenter of the lesion on PET images and on both T1-weighted images was measured, and misregistration was calculated. Statistical analysis was performed using Student t test. Misalignment values of the hepatic ablation zones between PET and MR images were calculated at 4.94 ± 1.35 mm (reader 1) and 4.89 ± 2.21 mm (reader 2) for Cartesian 3D-GRE sequence, and 2.48 ± 0.65 mm (reader 1) and 2.72 ± 0.44 mm (reader 2) for the radial 3D-GRE sequence, with p values of 0.0011 and 0.0133, respectively. Radial 3D-GRE offers improved registration accuracy with PET, supporting the use of this T1-weighted sequence in upper abdominal MR–PET studies.

Background MR reconstruction of virtual 3 dimensional volumes are routinely used to demonstrate cardiac malformations, particularly the complex geometric relationships between the great arteries and cardiac chambers. However, because of the need to signal average over multiple heart beats, rendering fast moving structures such as atrioventricular (AV) and semilunar valves has been limited. Echocardiography has sufficient frame rates to acquire images of moving valves in real-time 3d volumes.

International audience; We conducted a preclinical assessment on young macaques aimed at detecting white matter lesions. We present the protocol we implemented to achieve the lesions detection using a bedside non-invasive optical-based Time-Resolved instrumentation we have optimized for this purpose. We validated the reconstructed 3D absorption map with co-registration of MRI data. © (2015) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.

Diffusion tensor imaging (DTI) can provide important macroscopic structural information for mouse brain but is limited by the available imaging resolution and inferentia tractography. To validate DTI tractography, it is paramount to merge DTI images and microscopic images with true representation of fibers. Recently, we have developed a micro-optical sectioning tomography (MOST) system which enables neurite level resolution. By using Advanced Normalization Tools, we have aligned our MOST dataset to the template dataset of DTI from Duke University. To optimize the computing efficacy, lower resolution of two different modal images has been used to get a displacement field of diffeomorphism. Then the displacement field is extended to the deformation of full resolution images. This has been a flexible strategy for 3D nonlinear mouse brain registration across different modalities and different resolutions. Under this kind of co-registration, we can show very fine neural fiber architectures as foreground (MOST) on the background (DTI) of the fibre density map. By comparing the derived maps between DTI and MOST, we have realized that DTI technique encounters pitfall for resolving complex neural fibres. Image fusion of mouse DTI and MOST dataset should promote both neural circuits study and DTI applications.

Detection and matching of features from satellite images taken from different sensors, viewpoints, or at different times are important tasks when manipulating and processing remote sensing data for many applications. This paper presents a scheme for satellite image co-registration using invariant local features. Different corner and scale based feature detectors have been tested during the keypoint extraction, descriptor construction and matching processes. The framework suggests a sub-sampling process which controls the number of extracted key points for a real time processing and for minimizing the hardware requirements. After getting the pairwise matches between the input images, a full registration process is followed by applying bundle adjustment and image warping then compositing the registered version. Harris and GFTT have recorded good results with ASTER images while both with SURF give the most stable performance on optical images in terms of better inliers ratios and running time compared to the other detectors. SIFT detector has recorded the best inliers ratios on TerraSAR-X data while it still has a weak performance with other optical images like Rapid-Eye and ASTER.

Interferometric Synthetic Aperture Radar (InSAR) is a new measurement technology, making use of the phase information contained in Synthetic Aperture Radar (SAR) images. InSAR has been recognised as a potential tool for the generation of digital elevation models (DEMs) and the measurement of ground surface deformation. In InSAR data processing for the generation of DEMs, the first step is image co-registration, which brings both images into the same coordinate system for the generation of an interferogram for further processing. The accuracy of the resultant DEM will certainly be affected by the quality of the co-registration, which is in turn affected by the set of tie points used and some other factors. In this study, only the interval of tie points is considered. This means that other factors are kept unchanged. Four pairs of SAR images with size of 1760 × 400 pixels were used for testing. The effects are assessed by a relative measure for the quality of the resultant interferogram and an absolute measure for the accuracy of the resultant DEM. Results show that the effect of tie point interval on the accuracy of the final DEM is not linear. When the tie point interval is smaller than 273 × 44 pixels (273 pixels in row and 44 pixels in column), the variation in the resultant DEM accuracy is not significant. It is also noticeable that an interval of 205 × 34 pixels always results in the best or very good results. Therefore, it seems that an interval of around 200 × 30 pixels (200 pixels in row and 30 pixels in column) is an appropriate choice for the selection of tie points for image co-registration in hilly areas. Resume L'interferometrie par radar a synthese d'ouverture (InSAR) est une nouvelle technique de mesure qui utilise l'information sur la phase contenue dans les images issues de ces radars. On a depuis reconnu les possibilites qu'offrait l'interferometrie InSAR pour la creation des modeles numeriques des altitudes (MNA) et la determination des deformations superficielles du terrain. La premiere etape du traitement des donnees InSAR pour obtenir un MNA consiste a superposer les images en les amenant toutes deux dans le meme systeme de coordonnees, d'ou pourront resulter un interferogramme et son traitement ulterieur. La precision du MNA resultant est evidemment tributaire de la qualite de cette mise en superposition, qui depend a son tour de divers facteurs dont la qualite du jeu de points de liaison. Dans cette etude on n'a examine que l'influence de l'intervalle entre les points de liaison. On a donc suppose constants tous les autres facteurs. On a effectue un essai en utilisant quatre couples d'images de radar a synthese d'ouverture, comprenant 1760 × 400 pixels. On a evalue l'influence de cet intervalle de maniere relative d'apres la qualite de l'interferogramme resultant et de maniere absolue d'apres la precision du MNA final. Les resultats montrent que l'influence de cet intervalle entre points de liaison sur la precision du MNA resultant n'est pas lineaire. Lorsque cet intervalle est inferieur a 273 × 44 pixels (c'est-a-dire 273 pixels sur les lignes et 44 sur les colonnes) la variation de precision sur le MNA final n'est pas significative. Il convient de noter que l'on a obtenu d'excellents resultats, si ce n'est le meilleur, avec un intervalle de 205 × 34 pixels. En consequence, il semble tout a fait approprie de se fixer un intervalle d'environ 200 × 30 pixels (200 pixels en ligne et 30 en colonne) dans le choix des points de liaison pour superposer des images en terrain accidente. Zusammenfassung Interferometric Synthetic Aperture Radar (InSAR) ist eine neue Messtechnologie, bei der die Messungen auf die Phaseninformation beruhen, die in Bildern aus Synthetic Aperture Radar (SAR) enthalten ist. InSAR wird als mogliches Werkzeug zur Generierung Digitaler Hohenmodelle (DEMs) und zur Messung von Deformationen der Erdoberflache angesehen. Zur Ableitung Digitaler Hohenmodelle aus InSAR Daten ist als erster Schritt eine Registrierung beider Bilder in das gleiche Koordinatensystem erforderlich, um ein Interferogramm zu erhalten, das als Grundlage fur die weitere Auswertung verwendet wird. Die Genauigkeit des resultierenden Hohenmodells wird sicherlich durch die Qualitat der Registrierung beeinflusst, die selbst durch die Verteilung und die Zahl der Verknupfungspunkte und von weiteren Faktoren beeinflusst wird. In dieser Studie wird nur der Abstand von Verknupfungspunkten betrachtet, d.h. dass andere Faktoren unverandert bleiben. Vier Paare von SAR Bildern der Grose 1760 × 400 Pixel wurden fur die Tests benutzt. Die Einflusse wurden durch ein relatives Mas fur die Qualitat des abgeleiteten Interferogramms und ein absolutes Mas fur die Genauigkeit des abgeleiteten Digitalen Hohenmodells abgeschatzt. Die Ergebnisse zeigen, dass die Einflusse des Intervalls der Verknupfungspunkte auf die Genauigkeit des abgeleiteten Hohenmodells nicht linear ist. Wenn das Intervall kleiner als 273 × 44 Pixel ist (273 Pixel in der Zeile und 44 Pixel in der Spalte), sind die Variationen im abgeleiteten Hohenmodell nicht signifikant. Ein Intervall von 205 × 34 Pixel zeigte immer das beste oder sehr gute Ergebnisse. Daraus lasst sich schliesen, dass ein Intervall von ca. 200 × 30 Pixel (200 Pixel in der Zeile und 30 Pixel in der Spalte) eine geeignete Grose fur die Auswahl von Verknupfungspunkten zur Registrierung bei hugeligem Gelande darstellt. Resumen El Radar Interferometrico de apertura sintetica (InSAR) es una nueva tecnologia que aprovecha la informacion de fase contenida en las imagenes de radar de apertura sintetica (SAR). Sin duda, InSAR es una herramienta con mucho potencial tanto para la obtencion de modelos digitales de elevacion (MDE) como para la medida de la deformacion del terreno. El primer paso en el procesamiento de datos InSAR para la obtencion de un MDE es el corregistro de las imagenes con el fin de situar las dos imagenes en el mismo sistema de coordenadas y generar el interferograma que se procesa a continuacion. Indudablemente, la exactitud del MDE resultante se vera afectada por la calidad del corregistro que, a su vez, depende del conjunto de puntos de enlace utilizado y de algunos otros factores. Este estudio se centra en el intervalo de los puntos de enlace, manteniendo los otros factores sin cambios. El experimento se realiza con cuatro pares de imagenes SAR de 1760 × 400 pixeles. Los efectos se evaluan con una medida relativa de calidad del interferograma resultante y una medida absoluta de la exactitud del MDE obtenido. Los resultados senalan que el efecto del intervalo de los puntos de enlace en la exactitud del MDE final no es lineal. Cuando el intervalo entre los puntos es menor de 273 × 44 pixeles (273 pixeles en el eje y, 44 pixeles en el eje x), la variacion en la exactitud del modelo obtenido no es significativa. Con un intervalo de 205 × 34 pixeles siempre se obtienen los mejores o, al menos, muy buenos resultados. Por lo tanto, parece que un intervalo entorno a 200 × 30 pixeles es apropiado para la seleccion de puntos de enlace para realizar el corregistro en areas con predominio de colinas.

This paper presents a review of 3D reconstruction of coronary arteries based on the co-registration of Intravascular Ultrasound (IVUS) and coronary angiogram. This method has been proposed to ease the process of diagnosis and treatment for coronary artery disease, which is highlighted as a global disease. The purpose of this paper is to provide an introduction and understanding of the works that have been done. This paper consists of an introduction, importance of 3D coronary artery, registration process, background of IVUS and coronary angiogram, and a brief description about previous works.

Purpose To investigate the visualisation of novel external fiducial skin markers in photography and MRI. To co-register photographs and MR images, and additionally assess the spatial accuracy of these co-registrations with the view of future application in the investigation of forensically relevant soft tissue lesions. Methods and materials Strand-shaped fiducial markers were secured externally over hematomas on the thigh of 10 volunteers. The region of interest was photographed and examined using MRI at 3T in oblique and transversal orientations and the visibility of the markers assessed. Markers provided ‘control points' in both sets of images, enabling the computation of an affine transform to register oblique MR images to photographs. The fiducial registration error was evaluated by calculating the root-mean-square error of nine corresponding evaluation points visible in both modalities. Results Fiducial markers were clearly visualised in both photography and MRI. The co-registration of photographs and oblique MR images was achieved for all participants. The overall root-mean-square error for registrations was 1.18mm (TIRM) and 1.46mm (TSE2D with SPAIR fat-suppression). Conclusions The proposed approach led to the successful visualisation of non-invasive fiducial markers using photography and MRI (TIRM and TSE2D (SPAIR) sequences). This visualisation, combined with an affine transformation process provided a simple, cost-effective way to accurately co-register photographs and MR images of subcutaneous hematomas located on the thigh. Further investigation of the novel markers and the proposed co-visualisation approach holds potential to improve not only the forensic documentation of soft tissue lesions, but to also improve certain clinical applications, including the area of dermatology.

INTRODUCTION: Precise co-registration of brain tissue and medical imaging is critical for validation of novel imaging biomarkers meant to detect infiltrative brain cancer. Brain tissue distortion during fixation, and brain slicing in sub-optimal orientation can complicate co-registration. METHODS: Three high-grade glioma patients undergoing brain only autopsies were included in this analysis. A clinically acquired MRI was used to render 3D computer assisted drafting (CAD) models. To generate a fixation cage for preventing tissue distortion, a high-resolution 3D-T1-weighted MRI scan was used to generate a solidified mesh with approximately 500 holes to allow formalin exposure. The left and right hemispheric meshes were then 3D printed in plastic using a MakerBot Replicator2X. The brains were fixed in formalin for approximately 14 days within the plastic meshes. To generate a slicing mold for one patient, the shell from the 3D-T1-weighted scan was combined with the slice profile of 6.5mm thicker-cut images. A slicing mold was then designed and 3D-printed with slots oriented vertically with the brain lying on its side within the mold. To assess consistency, the thickness of each slice was measured. To assess the co-registration with the patient's clinical MRI, gyri and sulci from each MRI slice were identified and compared to the corresponding tissue slice. RESULTS: The 3D-printed brain cages took approximately 16hrs per hemisphere to print using approximately 1.6lbs of plastic (∼$35/brain). The custom brain-slicing mold took approximately 22hrs to print totaling about 1.2lbs of plastic (∼$26). The average tissue slice thickness was 6.9mm compared to the 6.5mm MRI slices. Of the 188 gyri and sulci identified on in tissue slices 73% co-localized precisely with those seen in corresponding MRI slices. CONCLUSION: 3D printed brain cages and slicing molds provide affordable means for preventing tissue distortion and precisely slicing brain tissue for co-registration with imaging acquired prior to death.

The problem of interferometric SAR (InSAR) image co‐registration is addressed. Classical InSAR matching considers only the translation shift between two complex images. However, for accurate radar image matching (different swaths, paths, or aspects angles), other shift parameters would be useful. For two‐dimensional images, the application of the Symmetric Phase Only Matching Filtering (SPOMF) to the Fourier–Mellin Invariant (FMI) descriptors allows an accurate and efficient registration of translated, rotated and scaled images. This paper discusses the utility and feasibility of the SPOMF technique on the FMI descriptors of two interferometric SAR images. The FMI‐SPOMF technique is applied on different pairs of InSAR data corresponding to areas with different relief types, in France, Nepal and Tunisia. The results are compared with those of classical cross‐correlation registration techniques. We find that the FMI‐SPOMF descriptors of an interferometric pair of images allows more precise co‐registration a...

We examined image co-registration techniques based on mutual information (MI) and normalized mutual information (NMI) to registrate two SPECT images with localized differences in brain activity. Simulations were performed by 100 trials with the random initial mismatch of ±20° and ±45 mm for a normal perfusion model (NRM) and a pathological perfusion model with a hemisphere defect (HSD) . In the HSD, activity in the defect region was reduced to be 75% (HSD75%), 50% (HSD50%), and 25% (HSD25%) of normal value. In the MI-based registration for the NRM and NRM dataset, no biases were observed (≤-0.06°, ≤0.06 mm) and the SDs were very small (≤0.07°, ≤0.04 mm) . In the registration for the NRM and HSD dataset, the biases were slightly larger (≤-0.18°, ≤0.34 mm for HSD75%, ≤-0.23°, ≤=0.45 mm for HSD50%, ≤-0.23°, ≤0.18 mm for HSD25%) in comparison with the results of registration for the NRM and NRM dataset. The accuracy of the NMI was almost identical to that of the MI. The entropy-based registration techniques are relatively unaffected by localized differences in brain activity.

Positron Emission Tomography (PET) with 18 F-FDG is a promising tool for the detection and evaluation of active inflammation in animal models of neuroinflammation. MRI is a complementary imaging technique with high resolution and contrast suitable to obtain the anatomical data required to analyze PET data. To combine PET and MRI modalities, we developed a support bed system compatible for both scanners that allowed to perform imaging exams without animal repositioning. With this approach, MRI and PET data were acquired in mice with Experimental autoimmune encephalomyelitis (EAE). In this model, it was possible to measure a variation of 18 F-FDG uptake proportional to the degree of disease severity which is mainly related to Central Nervous System (CNS) inflammation. Against the low resolved PET images, the co-registered MRI/PET images allowed to distinguish the different brain structures and to obtain a more accurate tracer evaluation. This is essential in particular for brain regions whose size is of the order of the spatial resolution of PET.

The aim of the present study was to access and compare the accuracy and reproducibility of two fully automatic and two manual methods for co-registration of structural T1-weighted MR images and functional images with localized binding. It is concluded that for functional images with distributed binding all over the brain, such as F18-FDG images, the automatic methods are preferable, while for images with localized binding, such as F18-Altanserin 5-HT 2A neuroreceptor images, it is essential to use methods that handle this special problem, like the proposed manual methods.

Objective: The Schaltenbrand-Wahren atlas (SWA) is anatomical, sparse and inconsistent in three dimensions. A high-resolution, electrophysiology-based atlas derived from numerous specimens overcomes these limitations. A combined anatomy-physiology atlas leverages the strengths and complementarity of both, which is studied here. Method: An electronic version of the SWA was constructed. A probabilistic functional atlas (PFA) was developed from electrophysiological and neuroimaging data with 0.25 mm3 resolution. A combined atlas is constructed by co-registering the PFA with the SWA by applying linear scaling along the intercommissural distance and the height of the thalamus. Results: An anatomy-physiology atlas is superior to its component atlases. When in register, the anatomical and functional atlases are displayed together and used simultaneously. A structure of interest from the PFA can be displayed with high resolution, compensating for its sparseness in the SWA, while its surrounding structures are obtained from the SWA. The best anatomical and functional targets can be compared: for the subthalamic nucleus, the horizontal and lateral coordinates of the SWA-based and PFA- based targets are approximately the same, while the latter is located 1.3–1.5 mm more anteriorly. Conclusion: An anatomy-physiology atlas may enhance the accuracy of targeting and increase the neurosurgeon’s confidence. It also opens new research avenues to serve as a reference for: (1) constructing an extendable atlas by adding new electrophysiological data, (2) comparing anatomical and functional targets, (3) studying and comparing various symptoms, and (4) comparing image-based targets obtained from various modalities.

aDepartment of Neuroradiology, University of Erlangen-Nurnberg, Schwabachanlage 6, D-91054, Germany bDepartment of Radiology, Division of Molecular and Gender Imaging, Medical University Vienna, Wahringer Gurtel 18–20, A-1090, Vienna, Austria c Karlsruhe Institute of Technology (KIT), Institute for Data Processing and Electronics, Postfach 3640, D-76021 Karlsruhe, Germany d Institute of Diagnostic and Interventional Radiology I, Friedrich-Schiller-University Jena, Erlanger Allee 101, D-07740 Jena, Germany

We investigated a co-registration algorithm using a contour-fitting procedure to integrate functional data from magnetoencephalography (MEG) and functional magnetic resonance imaging (fMRI) for frameless stereotaxy. In fMRI the shape of the head was reconstructed from anatomical images, in MEG it was scanned using an electromagnetic sensor position indicator. Functional information was transferred to the 3D-MR image set used for frameless stereotaxy by fitting the digitized (MEG) and reconstructed head shape (fMRI) to the 3D-MR images. The mean residual error of the contour fit was 2.3 mm for the MEG and 1.3 mm for the fMRI registration. According to computer simulations, the achievable transformation error is 0.75 and 0.5 mm, respectively. This method enables independent recording of functional and anatomical measurements with a co-registration accuracy better than 2 mm.

This paper describes a spatio-temporal registration approach for speech articulation data obtained from electromagnetic articulography (EMA) and real-time Magnetic Resonance Imaging (rtMRI). This is motivated by the potential for combining the complementary advantages of both types of data. The registration method is validated on EMA and rtMRI datasets obtained at different times, but using the same stimuli. The aligned corpus offers the advantages of high temporal resolution (from EMA) and a complete mid-sagittal view (from rtMRI). The co-registration also yields optimum placement of EMA sensors as articulatory landmarks on the magnetic resonance images, thus providing richer spatio-temporal information about articulatory dynamics. (C) 2014 Acoustical Society of America

Intravascular optical coherence tomography (IVOCT) has become the imaging modality of choice for the evaluation of coronary artery disease and percutaneous coronary intervention (PCI). Both for clinical practice and research, there is a growing interest in 3-dimensional (3D) visualization, as this gives a more comprehensive and intuitively easier to understand representation, compared with 2-dimensional, cross-sectional images. Integrating 3D-IVOCT with classic X-ray angiographic images offers additional advantages and the prospect of integrating IVOCT in fluoroscopic guidance during PCI. Different vendors of IVOCT technology already provide integrated 3D rendering software in their consoles, making 3D images available at the ‘push-of-a-button’. In this review, we will discuss (1) the basic principles and elaboration of 3D-IVOCT in recent years, (2) the feasibility and potential advantages of co-registration with X-ray angiography, (3) the currently available solutions for 3D imaging and their potential clinical applications, and (4) the ongoing development of applications for advanced 3D visualization.

Purpose 4D co-registration of X-ray- and MR-mammograms (XM and MM) is a new method of image fusion. The present study aims to evaluate its clinical feasibility, radiological accuracy, and potential clinical value. Methods XM and MM of 25 patients were co-registered. Results were evaluated by a blinded reader. Results Precision of the 4D co-registration was “very good” (mean-score [ms]=7), and lesions were “easier to delineate” (ms=5). In 88.8%, “relevant additional diagnostic information” was present, accounting for a more “confident diagnosis” in 76% (ms=5). Conclusion 4D co-registration is feasible, accurate, and of potential clinical value.

Background: Repeat Gamma Knife stereotactic radiosurgery (GKSR) procedures are becoming common, especially for brain metastases. It is important to identify tumors requiring treatment at repeat GKSR and it can be challenging to distinguish treated tumors, tumor progression and new tumors. Using the image co-registration tool within the Leksell Gamma Plan software, we developed a technique to aid in the identification of tumors needing treatment. Objectives: The objective was to explore a new co-registration technique to identify tumors requiring treatment at repeat GKSR procedures. Methods: Ten patients who underwent repeat GKSR for brain metastases were identified. Contrast-enhanced volumetric T1 magnetic resonance images (MRI) from the previous GKSR were co-registered with the new images and the resulting two-color format image was used to evaluate tumor status. Results: Using the co-registered images, tumors were characterized as: resolved, regressed, stable, larger or new. Overall, 13.6% of tumors completely resolved, 26.2% regressed, 13.1% remained stable, while 7.9% progressed. Thirty-nine percent of tumors were new. Conclusions: The co-registration technique makes clinically relevant changes conspicuous on MRI. It distinguishes between tumors potentially requiring treatment and those that have been treated successfully. It can be used with tumors other than metastases and for evaluating tumor response at follow-up.

Similar to blood oxygenation level-dependent (BOLD) functional magnetic resonance imaging (fMRI), functional nearinfrared spectroscopy (fNIRS) observes regional hemodynamic responses associated with neuronal activation. However, the conventional criteria for detecting true positive fNIRS and fMRI signals appear to be based on different understandings of cerebral hemodynamics. Considerable numbers of fNIRS studies have ascribed the increase in oxygenated hemoglobin to a typical sign of functional activation, whereas the corresponding BOLD signal in fMRI directly correlates with a decrease in deoxygenated hemoglobin. This inconsistency requires solution through the simultaneous measurements of fNIRS and fMRI. In practice, however, there remain several technical problems associated with conducting simultaneous measurements with high reproducibility. One issue is the precise spatial registration of NIRS optodes in MR images. We prepared marker containers of an annular shape that can be coaxially fixed to the optode. Liquid paraffin with α-tocopheryl acetate, which exhibits a bright contrast in T1-weighted MR images of human heads, was solidified in each container by adding higher fatty acid. A subject wearing the marker-fixed optodes at parietal area participated in preliminary fNIRS and fMRI experiments; the subject was instructed to execute single-sided hand finger tapping. The positions showed that deoxygenated hemoglobin decreases in fNIRS coincided with the BOLD-positive region in fMRI. The prepared marker is chemically stable and repetitively usable. We believe that this simple method contributes precision to the co-registration of fNIRS and fMRI.

Interpreting complex paediatric body MRI studies requires the integration of information from multiple sequences. Image processing software, some freely available, allows the radiologist to use simple and rapid post-processing techniques that may aid diagnosis. We demonstrate the use of fusion and subtraction post-processing techniques with examples from four areas of application: enterography, oncological imaging, musculoskeletal imaging and MR fistulography.

Image co-registration is a widely used processing in many application fields, such as medicine, computer vision and remote sensing. In each domain, different techniques may be applied, depending on the nature of the data avaliable and the wanted results. In this paper, we focused on optical and SAR based satellite images, that need to be co-registered for a variety of reasons, from change detection to interferometry, from agriculture monitoring to surveillance and so on. One of the main characteristic of satellite based imagery is the large amount of data that needs to be processed, that usually leads to high running time, thus limiting the kind of processing that can be effectively carried out. To solve this problem, a complete toolchain for image co-registration based on a GPU parallel architecture is presented: the tool flow, the detailed implementation, and experimental results are illustrated, showing the level of performance and accuracy achieved.

Accurate co-registration of MRI and EEG data is indispensable for the correct interpretation of EEG maps or source localizations in relation to brain anatomy derived from MRI. In this study, a method for the co-registration of EEG and MRI data is presented. The method consists of an iterative matching of EEG-electrode based reconstructions of the scalp surface to scalp-segmented MRIs. EEG-electrode based surface reconstruction is achieved via spline interpolation of individually digitized 3D-electrode coordinates. In contrast to other approaches, neither fiducial determination nor any additional provisions (such as bite bars, other co-registration devices or head shape digitization) are required, and co-registration errors associated with inaccurate fiducial determination are avoided. The accuracy of the method was estimated by calculating the root-mean-square (RMS) deviation of spline interpolated and MRI-segmented surface reconstructions in 20 subjects. In addition, the distance between co-registered and genuine electrode coordinates was assessed via a simulation study, in which surface reconstruction was based on virtual electrodes determined on the scalp surface of a high-resolution MRI data set. The mean RMS deviation of surface reconstructions was 2.43 mm, and the maximal distance between any two matched surface points was 5.06 mm. The simulated co-registration revealed a mean deviation of genuine and co-registered electrode coordinates of 0.61 mm. It is concluded that surface matching using spline interpolated reconstructions of scalp surfaces is a precise and highly practicable method to co-register EEG and MRI data.

Summary: A two-scale theoretical description outlines relationships between brain current sources and the resulting extracranial electric field ,r ecorded as EEG. Finding unknown sources of EEG, the so-called "inverse problem", is discussed in general terms, with emphasis on the fundamental non-uniqueness of inverse solutions. Hemodynamic signatures, measured with fMRI, are expressed as voxel integrals to facilitate comparisons with EEG. Two generally distinct cell groups (1 and 2), generating EEG and fMRI signals respectively, are embedded within the much broader class of synaptic action fields. Cell groups 1 and 2 may or may not overlap in specific experiments. Implications of this incomplete overlap for co-registration studies are considered. Each experimental measure of brain function is generally sensitive to a different kind of source activity and to different spatial and temporal scales. Failure to appreciate such distinctions can exacerbate conflicting views of brain function that emphasize either global integration or functional localization.

We present a method for the coregistration and topographic mapping of trans-cranial magnetic stimulation (TCMS) data on surface rendered images of the cortex, derived from Magnetic Resonance Images (MRI). We describe the TCMS procedure and the methods used to locate the TCM stimulation sites in the MRI coordinate system, and the algorithms needed to depict the TCMS distribution as a pseudocolour contour map on the cortical surface. The methods are validated using TCMS data from the hand (thenar) and leg (tibialis muscle). The methods used correctly depict the expected motor representations of each of these areas and we therefore propose that this technique may be used as a functional imaging tool in the investigation of cortical function in both normals and patients.

In image-guided cardiac interventions, X-ray imaging and intravascular ultrasound (IVUS) imaging are two often used modalities. Interventional X-ray images, including angiography and fluoroscopy, are used to assess the lumen of the coronary arteries and to monitor devices in real time. IVUS provides rich intravascular information, such as vessel wall composition, plaque, and stent expansions, but lacks spatial orientations. Since the two imaging modalities are complementary to each other, it is highly desirable to co-register the two modalities to provide a comprehensive picture of the coronaries for interventional cardiologists. In this paper, we present a solution for co-registering 2-D angiography and IVUS through image-based device tracking. The presented framework includes learning-based vessel detection and device detections, model-based tracking, and geodesic distance-based registration. The system first interactively detects the coronary branch under investigation in a reference angiography image. During the pullback of the IVUS transducers, the system acquires both ECG-triggered fluoroscopy and IVUS images, and automatically tracks the position of the medical devices in fluoroscopy. The localization of tracked IVUS transducers and guiding catheter tips is used to associate an IVUS imaging plane to a corresponding location on the vessel branch under investigation. The presented image-based solution can be conveniently integrated into existing cardiology workflow. The system is validated with a set of clinical cases, and achieves good accuracy and robustness.

Emotional stimuli are preferentially processed over neutral stimuli. Previous studies, however, disagree on whether emotional stimuli capture attention preattentively or whether the processing advantage is dependent on allocation of attention. The present study investigated attention and emotion processes by measuring brain responses related to eye movement events while 11 participants viewed images selected from the International Affective Picture System (IAPS). Brain responses to emotional stimuli were compared between serial and parallel presentation. An ‘emotional’ set included one image with high positive or negative valence among neutral images. A ‘neutral’ set comprised four neutral images. The participants were asked to indicate which picture - if any - was emotional and to rate that picture on valence and arousal. In the serial condition, the event-related potentials (ERPs) were time-locked to the stimulus onset. In the parallel condition, the ERPs were time-locked to the first eye entry on an image. The eye movement results showed facilitated processing of emotional, especially unpleasant information. The EEG results in both presentation conditions showed that the LPP (‘late positive potential’) amplitudes at 400–500 ms were enlarged for the unpleasant and pleasant pictures as compared to neutral pictures. Moreover, the unpleasant scenes elicited stronger responses than pleasant scenes. The ERP results, however, did not support parafoveal emotional processing, although the eye movement results suggested faster attention capture by emotional stimuli. Our findings, thus, suggested that emotional processing depends on overt attentional resources engaged in the processing of emotional content. The results also indicate that brain responses to emotional images can be analyzed time-locked to eye movement events, although the response amplitudes were larger during serial presentation.

Local hemodynamic forces, such as wall shear stress, are thought to trigger cellular and molecular mechanisms that determine atherosclerotic plaque vulnerability to rupture. Magnetic resonance imaging (MRI) has emerged as a powerful tool to characterize human carotid atherosclerotic plaque composition and morphology, and to identify plaque features shown to be key determinants of plaque vulnerability. Image-based computational fluid dynamics (CFD) has allowed researchers to obtain time-resolved wall shear stress (WSS) information of atherosclerotic carotid arteries. A deeper understanding of the mechanisms of initiation and progression of atherosclerosis can be obtained through the comparison of WSS and plaque composition and morphology. To date, however, advance in knowledge has been limited greatly due to the lack of a reliable infrastructure to perform such analysis. The aim of this study is to establish a framework that will allow for the co-registration and analysis of the three-dimensional (3D) distribution ofWSS and plaque components and morphology. The use of this framework will lead to future studies targeted to determining the role of WSS in atherosclerotic plaque progression and vulnerability.

Purpose To automatically co-register optical coherence tomography (OCT), colour fundus photography (CFP), red-free (RF) and fluorescein angiography (FA). Methods The burden of manually assisted co-registration and the number of images taken per day prevent its widespread use. An algorithm able to co-register images in between the major modalities in due time is herewith proposed. Typically, OCT fundus references translate into a poorly detailed vascular network, thus rendering difficult its co-registration to other imaging modalities. A recently developed method (by our group) to compute the vascular network from OCT to the level of detail of CFP was used. A set of vessel features is computed and an iterative process estimates the transformation required to co-register these locations. At each step the number of inliers is determined and the process repeats. Images of 20 eyes from 13 patients that underwent high-definition OCT, CFP and FA, were co-registered. These images were manually segmented by a grader. Two additional sets were co-registered and evaluated by an expert: 102 OCT/CFP image pairs from 51 subjects, and 40 FA/RF image pairs from 20 patients. Results A skeleton overlap metric was defined and computed based on the vessels skeleton (0–no overlap, 1-full overlap) for the first set. An average overlap of 0.93±0.03 (N=60) is reported. From the expert evaluation, 91% of images were successfully co-registered. For the registration of a given pair, the algorithm takes 1.68±0.38 seconds on an Intel Core i7-2600k, 3.4 GHz computer. Conclusion The achieved level of co-registration, render this process an asset to the clinical daily practice and research.

Bone metastases are a clinically significant problem that arises in approximately 70% of metastatic breast cancer patients. Once established in the bone, tumor cells induce changes in the bone microenvironment that lead to bone destruction, pain, and significant morbidity. While much is known about the later stages of bone disease, less is known about the earlier stages or the changes in protein expression in the tumor micro-environment. Due to promising results of combining magnetic resonance imaging (MRI) and Matrix-Assisted Laser Desorption/Ionization Imaging Mass Spectrometry (MALDI IMS) ion images in the brain, we developed methods for applying these modalities to models of tumor-induced bone disease in order to better understand the changes in protein expression that occur within the tumor-bone microenvironment. Specifically, we integrated 3-dimensional-volume reconstructions of spatially resolved MALDI IMS with high-resolution anatomical and diffusion weighted MRI data and histology in an intratibial model of breast tumor-induced bone disease. This approach enables us to analyze proteomic profiles from MALDI IMS data with corresponding in vivo imaging and ex vivo histology data. To the best of our knowledge, this is the first time that these three modalities have been rigorously registered in the bone. The MALDI mass-to-charge ratio peaks indicate differential expression of calcyclin, ubiquitin, and other proteins within the tumor cells, while peaks corresponding to hemoglobin A and calgranulin A provided molecular information that aided in the identification of areas rich in red and white blood cells, respectively. This multi-modality approach will allow us to comprehensively understand the bone-tumor microenvironment and thus may allow us to better develop and test approaches for inhibiting bone metastases.

Brain shift, the change in configuration of the brain after opening the dura mater, is a key problem in neuronavigation. We present an approach to co-register intra-operative microscope images with preoperative MRI data to adapt and optimize intra-operative neuronavigation. The tools are a robust classification of sulci on MRI extracted cortical surfaces, guided user marking of most prominent sulci on a microscope image, and the actual variational registration method with a fidelity energy for 3D deformations of the cortical surface combined with a higher order, linear elastica type prior energy. Furthermore, the actual registration is validated on an artificial testbed and on real data of a neuro clinical patient.

Purpose: To compare typical volumetric spatial distortions for 1.5 Tesla versus 3 Tesla MRI Gamma Knife radiosurgery scans in the frame marker fusion and co-registration frame-less modes. Methods: Quasar phantom by Modus Medical Devices Inc. with GRID image distortion software was used for measurements of volumetric distortions. 3D volumetric T1 weighted scans of the phantom were produced on 1.5 T Avanto and 3 T Skyra MRI Siemens scanners. The analysis was done two ways: for scans with localizer markers from the Leksell frame and relatively to the phantom only (simulated co-registration technique). The phantom grid contained a total of 2002 vertices or control points that were used in the assessment of volumetric geometric distortion for all scans. Results: Volumetric mean absolute spatial deviations relatively to the frame localizer markers for 1.5 and 3 Tesla machine were: 1.39 ± 0.15 and 1.63 ± 0.28 mm with max errors of 1.86 and 2.65 mm correspondingly. Mean 2D errors from the Gamma Plan were 0.3 and 1.0 mm. For simulated co-registration technique the volumetric mean absolute spatial deviations relatively to the phantom for 1.5 and 3 Tesla machine were: 0.36 ± 0.08 and 0.62 ± 0.13 mm with max errors of 0.57 and 1.22 mm correspondingly. Conclusion: Volumetric spatial distortions are lower for 1.5 Tesla versus 3 Tesla MRI machines localized with markers on frames and significantly lower for co-registration techniques with no frame localization. The results show the advantage of using co-registration technique for minimizing MRI volumetric spatial distortions which can be especially important for steep dose gradient fields typically used in Gamma Knife radiosurgery. Consultant for Elekta AB

An automated registration method for the correction of AATSR co-registration is presented. AATSR co-registration between the nadir and forward views is known to be poor, generally greater than 1 pixel. Here we apply the SURF algorithm to automatically generate a set of tie-points between the forward and nadir images. These tie-points are then used to create a set of warping parameters which effectively co-register the forward and nadir images. The warping parameters have been tested globally for a number of years of AATSR data and are shown to provide improved co-registration to less than a pixel in all cases. The same method will, in the future, be applied to the other ATSR instruments which share similar co-registration problems.

We evaluated our co-registration method which uses the number of SPs and amplitude information for ALOS data. Experimental results showed that our proposed method improves the quality of the interferogram as much as the case of JERS-1 data.

Aims: To demonstrate the feasibility and potential usefulness of an offline fusion of matched optical coherence tomography (OCT) and intravascular ultrasound (IVUS)/virtual histology (IVUS-VH) images. Methods and results: A total of nine matched OCT, IVUS, and IVUS-VH cross-sections were fused according to a methodology described in this report. On the basis of the fused images, an OCT-IVUS-VH tissue classification algorithm is proposed taking into account the individual strength of both techniques. Conclusions: Offline fusion of co-registered IVUS and OCT is feasible and combines the strengths of both imaging modalities, potentially improving the diagnostic accuracy of plaque characterisation.

Synthetic aperture radar (SAR) has great advantage of being able to observe large area accurately in any weather. It can measure various properties of the earth Boerner (2003), e.g., the topography, the vegetation Hajinsek et al. (2009), and the landscape changes caused by earthquakes or volcanoes Gabriel et al. (1989). One of the common usages of SAR is Interferometric synthetic aperture radar (InSAR), which can measure the landscape with an interferogram of SAR images. The SAR interferogram is made from two complex-valued maps obtained by observing identical place, named "master" and "slave." The phase information of the interferogram corresponds to the ground topography. To generate digital elevation model (DEM) from interferogram, phase unwrapping process is required. However, an unwrapping process is disturbed seriously by singular points (SPs), the rotational points existing in the phase map.

The correlation of multiple in vivo imaging modalities with graded whole-mount histology ideally requires co-registration with appropriately quantified accuracy. This work presents a framework for the co-registration of multiple in vivo prostate images with histology as a composition of pair wise registrations. This work also proposes a methodology for the measurement of the accuracy of these registrations based on a composition of target registration errors calculated from homologous landmarks in registered pairs of images. Preliminary results from the implementation of components of the framework are reported.