Search

This paper proposes a new algorithm based on energy features for noise reduction using wavelets. The device noise profile is obtained by the noise images taken from the imaging device so that it can represent the device's noise in multi-scale and multi-band. The energy feature takes advantage of the inter-scale relationship and spatial relationship of wavelet transformation. The wavelet coefficients are shrunk by the likelihood of noise or signal based on its energy level. The de-noised images are obtained by wavelet reconstruction. The results and comparison against common used methods show that the performance of our method is very promising despite simple structure.

Traditional neuropathological examination provides information about neurological disease or injury of a patient at a high-resolution level. Correlating this type of post mortemdiagnosis with in vivoimage data of the same patient acquired by non-invasive tomographic scans greatly complements the interpretation of any disease or injury. We present the validation of a registration method for correlating macroscopic pathological images with MR images of the same patient. This also allows for 3-D mapping of the distribution of pathological changes throughout the brain. As the validation deals with datasets of widely differing sampling, we propose a method using smooth curvilinear anatomical features in the brain which allows interpolation between wide-spaced samples. Curvilinear features are common anatomically, and if selected carefully have the potential to allow determination of the accuracy of co-registration across large areas of a volume of interest.

In this paper, we present a systematic investigation of the variability of the human cortical folding using a generic gyral model (GGM). The GGM consists of a fixed number of vertices that can be registered non-linearly to an individual anatomy so that for each individual we have a clearly defined set of landmarks that is spread across the cortex. This allows us to obtain a regionalized estimation of inter-subject variability. Since the GGM is stratified into different levels of depth, it also allows us to estimate variability as a function of depth. As another application of a polygonal line representation underlying the generic gyral model, we present a cortical parcellation scheme that can be used to regionalize cortical measurements.

Accurate co-registration of MR images of the brain with pathological data from the same patient is required for many applications, including validation of in-vivo MRI processing. We present preliminary results from a 2-D optical to 3-D MRI registration process, which depends on initialisation using an initial 3-D reconstruction of post-mortem whole-brain or hemisphere slices from independent 2D images followed by 3-D to 3-D registration. Our overall approach progressively models the acquisition process, and the final stages account for variations in slice thickness and other errors that can occur as a result of the specimen preparation and earlier processing. Our method can be used in most neuropathology laboratories where thick slices are cut. Alternative approaches depend on sectioning by macrocryotomes, which are not generally available and have limited applicability in conditions such as CJD where infectivity risks are high.

The deep grey matter structures of the brain have been reported to show MR hyperintensity in the sporadic form of Creutzfeldt-Jakob disease (sCJD), but the criteria for visual judgment of this are not well defined. We carried out a quantitative study of T2 weighted and proton density scans comparing 10 sCJD patients with 10 non-CJD dementia controls (NCD) and also with 11 patients suffering from the new variant form of CJD (vCJD). Scans were acquired in a clinical context and came from many hospitals. Absolute intensities varied widely and did not allow any useful discrimination. In all groups the putamen had a gradient of reducing intensity from anterior to posterior on T2 scans. In both s- and v- CJD patients this gradient was increased. Sensitivity and specificity (SS (2) intensities of deep grey matter structures vary systematically and intensity-based segmentation methods used in patients and normals should take account of this.

We present a method for registration of macroscopic optical images with MR images of the same patient. This forms a key part of a series of procedures to allow post mortem findings to be accurately registered with MR images, and more generally provides a method for 3D mapping of the distribution of pathological changes throughout the brain. The first stage of the method involves a 3D reconstruction of 2D brain slices and was presented in a previous paper [2]. In the current paper, we focus on the registration of the reconstructed volume with corresponding MR images.

Quantitative analysis of MR images is becoming increasingly important as a surrogate marker in clinical trials in multiple sclerosis (MS). This paper describes a fully automated model-based method for segmentation of MS lesions from multi-channel MR images. The method simultaneously corrects for MR field inhomogeneities, estimates tissue class distribution parameters and classifies the image voxels. MS lesions are detected as voxels that are not well explained by the model. The results of the automated method are compared with the lesions delineated by human experts, showing a significant total lesion load correlation and an average overall spatial correspondence similar to that between the experts.

This paper presents a new methodology for the automatic segmentation and characterization of object changes in time series of three-dimensional data sets. The purpose of the analysis is a detection and characterization of objects based on their dynamic changes. The technique was inspired by procedures developed for the analysis of functional MRI data sets. After precise registration of serial volume data sets to 4-D data, we applied a new time series analysis taking into account the characteristic time function of variable lesions. The images were preprocessed with a correction of image field inhomogeneities and a normalization of the brightness function over the whole time series. This leads to the hypothesis that static regions remain unchanged over time, whereas local changes in tissue characteristics cause typical functions in the voxel’s time series. A set of features are derived from the time series and their derivatives, expressing probabilities for membership to the sought structures. These multiple sources of uncertain evidence were combined to a single evidence value using Dempster Shafer’s theory. Individual processing of a series of 3-D data sets is therefore replaced by a fully 4-D processing. To explore the sensitivity of time information, active lesions are segmented solely based on time fluctuation, neglecting absolute intensity information.

The evaluation of spatial correspondence between binary objects resulting from a segmentation step performed by two different observers or methods is a critical part of the validation of a segmentation criterion or technique. Several global measures of correspondence have been previously proposed, but all of them assume a one-to-one correspondence between objects, thus failing to address local problems such as the splitting of an object by one of the observers. Moreover, such global measures do not distinguish between the reference and the observed objects and most of them lack a solid theoretical foundation. In this paper, we introduce a set of spatial correspondence indices that can evaluate global (many-to-many), local (many-to-one) and individual (one-to-one) spatial correspondence between observed and reference objects and vice versa. The proposed measures, derived from applying information theory concepts to the problem of spatial correspondence, are shown to be well-behaved and suitable to be used in medical imaging applications.

Segmentation of objects of interest in magnetic resonance imaging is a necessary procedure for volumetric calculations. However, these direct measurements tend to be inaccurate due to the intrinsic MRI partial volume effects. In this paper, a generalised method for correcting these effects based on the geometry and grey level intensity of the segmented objects is presented. This method is independent of the segmentation strategy used to extract the objects of interest. An evaluation is presented for three different segmentation methods and it is shown that the proposed generalised partial volume correction can improve the volume estimation of all three methods.

International audience; The VISLAN system is based on one of the newest developments in optical localiser tracking, where the localiser reflects light shone on it (passive illumination) rather than emitting light. In experiments with a CT-compatible skull phantom we evaluated the VISLAN system in comparison with a widely used stereotactic system, the Cosman-Roberts-Wells frame (Radionics lnc). The mean Euclidean difference in localising points initially defined in CT images was 2.5mm across a wide range of target positions. We estimated the accuracy of the VISLAN system with the same phantom set up but independently of the stereotactic frame, and showed a mean Euclidean error of 1.0mm. We concluded that the majority of the discrepancy between the two systems was accounted for by the inaccuracy of the CRW frame. It would be inappropriate to regard the frame based systems as a satisfactory gold standard for estimating the accuracy of VISLAN. However, frame-based and frameless guidance may be needed in the same operation, and it is important for such systems to be compared.

This paper describes registration updating for tracking patientcamera movement using marker pins in a video based neurosurgical guidance system(VISLAN). Both the algorithm and the experiment results of accuracy evaluation are given. The results show that the marker pin-based registration updating is a feasible solution for tracking patient-camera movement and has a very reasonable accuracy, particularly at spatial locations close to the markers pins, e.g. it has better than 0.7mm accuracy at distance of 50mm from the centre of the gravity of marker pins.

Many neuro- and ENT surgical procedures are performed using the operating microscope. Conventionally, the surgeon cannot accurately relate information from preoperative radiological images to the appearance of the surgical field. We propose that the best way do this is to superimpose image derived data upon the operative scene. We create a model of relevant structures (e.g. tumor volume, blood vessels and nerves) from multimodality preoperative images. By calibrating microscope optics, registering the patient in-theatre to image coordinates, and tracking the microscope intra-operatively, we can generate stereo projections of the 3D model and project them into the microscope eyepieces, allowing critical structures to be overlayed on the operative scene in the correct position. We have completed initial evaluation with a head phantom, and are about to start clinical evaluation on patients. With the head phantom a theoretical accuracy of 4.6mm was calculated and the observed accuracy ranged from 2mm to 5mm.

This paper describes two new methods of rendering multimodal images developed for a neurosurgical planning and guidance system (VISLAN). In our volume rendering technique we introduce a colour dependent filtering mechanism that enhances the representation of objects and improves the visualisation of spatial relationships. To achieve a good compromise between rendering speed and image quality, surface rendering is divided into two processes, a fast surface voxel projection and a surface refining and shading process. By considering the reflections from voxels both near to and on a surface in shading calculations, renderings become less sensitive to small surface extraction errors. A scheme which intermixes the volume rendering for some objects and surface rendering for others in the same scene is also presented. We show examples to illustrate each method in the context of preoperative surgical planning and intraoperative guidance.

We describe a technique for registering 3-D multimodal image data, acquired preoperatively, with intraoperative surface data derived from stereo video during neurosurgery. Ultimately, our aim is to provide a system that supplants traditional frame-based stereotactic techniques while achieving comparable accuracy. For registration we employ chamfer-matching in conjunction with a cost function that is robust to ‘outliers’. To balance robustness and computation speed, we employ a quasi-stochastic search of parameter space that includes pursuing multiple start points. This paper describes the registration problem as it pertains to our application. We discuss our approach to optimization and carry out a computational evaluation of the technique under various conditions.

Our previous work on knowledge based interpretation of angiographic images (Ritchings and Colchester 1986; Ritchings et al 1986; Wang et al 1987) has underlined the need to establish a hierarchical data structure which represents the image at multiple scales and which allows simplified control of the interaction between image-driven and knowledge-driven processes. Multi-resolution methods which use Gaussian blurring of different degrees in the analysis of features at multiple scales (Pizer et al 1987, 1988) have some theoretical advantages but, in practice, the selection of the most appropriate scales at which to smooth the image is difficult. We have developed a hierarchical algorithm which groups small elements into larger elements and then groups these into larger elements and so on. The size of the groups is determined by local criteria and prior decisions about scale are not required.

We propose a method of extracting and describing the shape of features from medical images which provides both a skeleton and boundary representation. This method does not require complete closed boundaries nor regularly sampled edge points. Lines between edge points are connected into boundary sections using a measure of proximity. Alternatively, or in addition, known connectivity between points (such as that available from traditional edge detectors) can be incorporated if known. The resultant descriptions are objectcentred and hierarchical in nature with an unambiguous mapping between skeleton and boundary sections.

This paper describes two methods to aid interpretation and quantification of SPECT or PET images. In the first method 3D SPECT or PET data sets are aligned and scaled to a 3D MRI data set of the same patient using 4 skin markers visible on each modality. Three display schemes have been implemented for viewing the aligned slices. Examples of these displays are provided. The second method uses a labelled 3D MRI reference data set from a volunteer to identify major anatomical structures. The MR reference data set is aligned with the isotope image using the same 4 markers plus a marker on the vertex of the skull. The reference data set is segmented approximately into the major tissue types - cerebrospinal fluid (CSF) and grey and white matter. Major structures are identified via labels in the 3D data set. A linked cursor aids delineation of anatomical regions on the isotope image using the outline of structures on the reference data set as a template. Directions for future research in the generation of complete digital anatomical atlases, which include inter-individual variations, are outlined.

The development of a computer model is described which will predict the concentration of contrast media (CM) in a blood vessel after the injection of CM and hence predict the angiographic appearances of the vessel. This model is used to calculate parametric images of CM concentration versus time and distance along the vessel for a wide range of experimental parameters. Blood velocity estimates are derived from these images at many points, both along the vessel and over the cardiac cycle. The model will be used to investigate the correct strategy for the injection of CM and to predict the precision of the resulting flow estimates.