Search

We present an advanced software tool designed for visualization and quantitative analysis of Diffusion Tensor Imaging (DTI) called Saturn. The software is specially developed to help clinicians and researchers in neuroimaging, and includes a complete set of visualization capabilities to browse and analyze efficiently DTI data, making this application a powerful tool also for diagnosis purposes. The software includes a robust quantification method for DTI data, using an atlas-based method to automatically obtain equivalent anatomical fiber bundles and regions of interest among different DTI data sets. Consequently, a set of measurements is also implemented to perform robust group studies among subjects affected by neurological disorders and control groups in order to look for significant differences. Finally, a comparison study with five similar DTI applications is presented, showing the advantages offered by this tool.

In this work, we propose to use fiber tracking in order to analyze and quantify the state of the pyramidal tracts in patients affected by tumors. We introduce a framework that includes an automatic method to obtain the fibers involved in the pyramidal tract of any subject, in order to compare robustly fiber bundles affected by tumors with healthy fiber tracts from control subjects and also to quantify the relative state of degeneration between the fiber tracts in the two hemispheres of the same patient. The comparative analyses proposed in our methodology are based on a new set of measures on the pyramidal tract, which take into account intrinsic properties of the fibers involved in the bundle as well as the similarity with the pyramidal tract of a standard healthy subject, modeled as the average of a set of controls. In order to perform better comparison studies and to take into account more information of the whole bundle, a mapping technique is used to represent the fiber tracts in 2D. Here, we show a set of experiments using 5 tumor patients and 10 control subjects, including pre- and post-operative studies in patients that have been treated with partial or total tumor resection. The results obtained indicate the usefulness of the method showing good overall performance. A reproducibility study using several acquisitions of the same patient is also presented to validate the techniques employed.

Discontinuation of mechanical ventilation is a challenging task that involves a number of subtle clinical issues. The gradual removal of the respiratory support (referred to as weaning) should be performed as soon as autonomous respiration can be sustained. However, the prediction rate of successful extubation is still below 25% based on previous studies. Construction of an automatic system that provides information on extubation readiness is thus desirable. Recent works have demonstrated that the breathing pattern variability is a useful extubation readiness indicator, with improving performance when multiple respiratory signals are jointly processed. However, the existing methods for predictor extraction present several drawbacks when length-limited time series are to be processed in heterogeneous groups of patients. In this paper, we propose a model-based methodology for automatic readiness prediction. It is intended to deal with multichannel, nonstationary, short records of the breathing pattern. Results on experimental data yield an 87.27% of successful readiness prediction, which is in line with the best figures reported in the literature. A comparative analysis shows that our methodology overcomes the shortcomings of so far proposed methods when applied to length-limited records on heterogeneous groups of patients.

In this paper, we focus on the problem of automatic noise parameter estimation for additive and multiplicative models and propose a simple and novel method to this end. Specifically we show that if the image to work with has a sufficiently great amount of low-variability areas (which turns out to be a typical feature in most images), the variance of noise (if additive) can be estimated as the mode of the distribution of local variances in the image and the coefficient of variation of noise (if multiplicative) can be estimated as the mode of the distribution of local estimates of the coefficient of variation. Additionally, a model for the sample variance distribution for an image plus noise is proposed and studied. Experiments show the goodness of the proposed method, specially in recursive or iterative filtering methods.

It has been shown that the tensor calculation is very sensitive to the presence of noise in the acquired images, yielding to very low quality Diffusion Tensor Images (DTI) data. Recent investigations have shown that the noise present in the Diffusion Weighted Images (DWI) causes bias effects on the DTI data which cannot be corrected if the noise characteristic is not taken into account. One possible solution is to increase the minimum number of acquired measurements (which is 7) to several tens (or even several hundreds). This has the disadvantage of increasing the acquisition time by one (or two) orders of magnitude, making the process inconvenient for a clinical setting. We here proposed a turn-around procedure for which the number of acquisitions is maintained but, the DWI data are filtered prior to determining the DTI. We show a significant reduction on the DTI bias by means of a simple and fast procedure which is based on linear filtering; well-known drawbacks of such filters are circumvented by means of anisotropic neighborhoods and sequential application of the filter itself. Information of the first order probability density function of the raw data, namely, the Rice distribution, is also included. Results are shown both for synthetic and real datasets. Some error measurements are determined in the synthetic experiments, showing how the proposed scheme is able to reduce them. It is worth noting a 50% increase in the linear component for real DTI data, meaning that the bias in the DTI is considerably reduced. A novel fiber smoothness measure is defined to evaluate the resulting tractography for real DWI data. Our findings show that after filtering, fibers are considerably smoother on the average. Execution times are very low as compared to other reported approaches which allows for a real-time implementation.

To eliminate the need of spatial intraframe regularization in a recently reported dynamic MRI compressed-sensing-based reconstruction method with motion compensation and to increase its performance.We propose a new regularization metric based on the introduction of a spatial weighting measure given by the Jacobian of the estimated deformations. It shows convenient discretization properties and, as a byproduct, it also provides a theoretical support to a result reported by others based on an intuitive design. The method has been applied to the reconstruction of both short and long axis views of the heart of four healthy volunteers. Quantitative image quality metrics as well as straightforward visual assessment are reported.Short and long axis reconstructions of cardiac cine MRI sequences have shown superior results than previously reported methods both in terms of quantitative metrics and of visual assessment. Fine details are better preserved due to the lack of additional intraframe regularization, with no significant image artifacts even for an acceleration factor of 12.The proposed Jacobian Weighted temporal Total Variation results in better reconstructions of highly undersampled cardiac cine MRI than previously proposed methods and sets a theoretical ground for forward and backward predictors used elsewhere. Magn Reson Med 77:1208-1215, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

Human body heat emission and others external causes can interfere in magnetic resonance image acquisition and produce noise. In this kind of images, the noise, when no signal is present, is Rayleigh distributed and its wavelet coefficients can be approximately modeled by a Gaussian distribution. Noiseless magnetic resonance images can be modeled by a Laplacian distribution in the wavelet domain. This paper proposes a new magnetic resonance image denoising method to solve this fact. This method performs shrinkage of wavelet coefficients based on the conditioned probability of being noise or detail. The parameters involved in this filtering approach are calculated by means of the expectation maximization (EM) method, which avoids the need to use an estimator of noise variance. The efficiency of the proposed filter is studied and compared with other important filtering techniques, such as Nowak’s, Donoho-Johnstone’s, Awate-Whitaker’s, and nonlocal means filters, in different 2D and 3D images.

In this paper a simple method for testing the equality of two time series is presented. It is intended as an easy-touse tool in chronobiological environments, though the procedure carries over directly to other applications. Simulation results support the conclusions drawn in the paper.

Several techniques have been described in the literature in recent years for the reconstruction of a regular volume out of a series of ultrasound (US) slices with arbitrary orientations, typically scanned by means of US freehand systems. However, a systematic approach to such a problem is still missing. This paper focuses on proposing a theoretical framework for the 3-D US volume reconstruction problem. We introduce a statistical method for the construction and trimming of the sampling grid where the reconstruction will be carried out. The results using in vivo US data demonstrate that the computed reconstruction grid that encloses the region-of-interest (ROI) is smaller than those obtained from other reconstruction methods in those cases where the scanning trajectory deviates from a pure straight line. In addition, an adaptive Gaussian interpolation technique is studied and compared with well-known interpolation methods that have been applied to the reconstruction problem in the past. We find that the proposed method numerically outperforms former proposals in several control studies; subjective visual results also support this conclusion and highlight some potential deficiencies of methods previously proposed.

This paper presents a common stochastic modelling framework for physiological signals which allows patient simulation following a synthesis-by-analysis approach. Within this framework, we propose a general model-based methodology able to reconstruct missing or artifacted signal intervals in cardiovascular monitoring applications. The proposed model consists of independent stages which provide high flexibility to incorporate signals of different nature in terms of shape, cross-correlation and variability. The reconstruction methodology is based on model sampling and selection based on a wide range of boundary conditions, which include prior information. Results on real data show how the proposed methodology fits the particular approaches presented so far for electrocardiogram (ECG) reconstruction and how a simple extension within the framework can significantly improve their performance.

Compressed sensing methods with motion estimation and compensation techniques have been proposed for the reconstruction of accelerated dynamic MRI. However, artifacts that naturally arise in compressed sensing reconstruction procedures hinder the estimation of motion from reconstructed images, especially at high acceleration factors. This work introduces a robust groupwise nonrigid motion estimation technique applied to the compressed sensing reconstruction of dynamic cardiac cine MRI sequences.A spatio-temporal regularized, groupwise, nonrigid registration method based on a B-splines deformation model and a least squares metric is used to estimate and to compensate the movement of the heart in breath-hold cine acquisitions and to obtain a quasistatic sequence with highly sparse representation in temporally transformed domains.Short axis in vivo datasets are used for validation, both original multicoil as well as DICOM data. Fully sampled data were retrospectively undersampled with various acceleration factors and reconstructions were compared with the two well-known methods k-t FOCUSS and MASTeR. The proposed method achieves higher signal to error ratio and structure similarity index for medium to high acceleration factors.Reconstruction methods based on groupwise registration show higher quality reconstructions for cardiac cine images than the pairwise counterparts tested.

Low contrast in ultrasound images caused by the granular pattern (speckle) makes difficult computational analysis and diagnosis. Thus, speckle filtering is a common step before computed automatic analysis. However, speckle depends on the inner echo-morphology of tissue and it should be removed without over-filtering relevant details for diagnostic purposes. Some methods were proposed to preserve important details by means of anisotropic diffusion schemes. However, they require to solve an evolutionary partial differential equation, which needs considerable computation time thatmakes this kind of filters impractical for real-time scenarios. Additionally, there is no rational criteria to select the optimal stop criteria. Some other detail preserving filters are based in the Non-Local means philosophy, however they involves an even higher computational cost.

In this paper, we propose a model–based methodology aimed at reconstructing corrupted or missing intervals of ECG signals acquired together with a PPG signal. To this end, we first estimate a joint–model of the ECG and PPG signals from the largest uncorrupted piece. Then, in a second stage, a set of candidates to replace the corrupted epoch is synthesized by sampling the aforementioned model. Each sample is evaluated respect to a boundary condition in order to select the best candidate. This signal is refined through an iterative method relying on the actual PPG data acquired during the corruption interval. Experiments on real data, show the capability of the proposed methodology to accurately reconstruct ECG pieces, outperforming so far presented solutions not accounting for joint information.

We propose a fully three-dimensional methodology for the computation of myocardial non-viable tissue transmurality in contrast enhanced magnetic resonance images. The outcome is a continuous map defined within the myocardium where not only current state-of-the-art measures of transmurality can be calculated, but also information on the location of non-viable tissue is preserved. The computation is done by means of a partial differential equation framework we have called Multi- Stencil Streamline Fast Marching (MSSFM). Using it, the myocardial and scarred tissue thickness is simultaneously computed. Experimental results show that the proposed 3D method allows for the computation of transmurality in myocardial regions where current 2D methods are not able to as conceived, and it also provides more robust and accurate results in situations where the assumptions on which current 2D methods are based —i.e., there is a visible endocardial contour and its corresponding epicardial points lie on the same slice—, are not met.

Detection and characterization of vulnerable plaques is of paramount importance for early disease diagnosis and treatment of carotid and coronary vascular accidents, which are mostly caused by this type of plaques. In this scope, the echo-morphology and composition are described in the literature, in statistical frameworks, by using several distributions to describe ultrasonic data depending on tissues, acquisition conditions, and equipment. Among them, the Rayleigh distribution and one of its generalization, the Nakagami distribution, have been extensively used to describe the raw envelope RF ultrasound signal. However, they fail to describe the statistical properties throughout some stages of the acquisition processes. Specifically, some signal-processing procedures usually used in the common ultrasound scanners involve linear filters or interpolations which affect the statistics of the signal and, eventually, change the assumed distribution of the signal. In this chapter, we provide an analysis of the statistical changes of the signal throughout the acquisition process of intravascular ultrasonography. The way interpolation and linear filtering procedures affect the statistics of the RF raw data receives here particular attention. In fact, we show that Gamma distribution is more suitable for statistical description of RF signal than the usual Rayleigh or Nakagami distributions, mainly because the interpolation and filtering procedures present in the US raw data processing pipeline. In this perspective, a Gamma mixture model (GMM) is proposed to describe the RF envelope images as natural evolution of a similar approach using Rayleigh distributions proposed before. This model is compared to the Rayleigh and Nagakami mixture models and the comparison results evidence a better performance with respect to plaque description and characterization of echogenic contents. As an application of this characterization, a probabilistic filter with detail preservation is proposed for detection and classification purposes of different type of tissues within the plaque. This tissue constitution information is the basis for plaque risk assessment and vascular accident prevention.

We propose the integration of biomechanical properties in a Markov random field of deformations for the estimation of myocardial motion in cine magnetic resonance images of patients affected by different cardiomyopathies.

The Ubiquitous Tele-monitoring Kit (UTK) is aimed at helping the clinician in acquiring, managing and normalizing information coming from the patients at home and storing the data in a distributed system. This approach aims to reduce costs in the healthcare system and alleviate the inherent problems to chronic patients who usually may visit the doctors. The proposed system closes the tele-monitoring loop; wearable and ergonomic sensors integrated in the textile, automatic detection and storage of the information coming from the sensors in a common platform (XNAT) and processing and visualization of the retrieved physiological signals presenting the data treated to the clinician to optimize the continuum of care and the decision-making.

Several statistical models have been proposed in the literature to describe the behavior of speckles. Among them, the Nakagami distribution has proven to very accurately characterize the speckle behavior in tissues. However, it fails when describing the heavier tails caused by the impulsive response of a speckle. The Generalized Gamma (GG) distribution (which also generalizes the Nakagami distribution) was proposed to overcome these limitations. Despite the advantages of the distribution in terms of goodness of fitting, its main drawback is the lack of a closed-form maximum likelihood (ML) estimates. Thus, the calculation of its parameters becomes difficult and not attractive. In this work, we propose (1) a simple but robust methodology to estimate the ML parameters of GG distributions and (2) a Generalized Gama Mixture Model (GGMM). These mixture models are of great value in ultrasound imaging when the received signal is characterized by a different nature of tissues. We show that a better speckle characterization is achieved when using GG and GGMM rather than other state-of-the-art distributions and mixture models. Results showed the better performance of the GG distribution in characterizing the speckle of blood and myocardial tissue in ultrasonic images.

International audience; Linear Minimum Mean Squared Error Estimation (LMMSE) is a simple, yet powerful denoising technique within MRI. It is based on the computation of the mean and variance of the data being filtered according to a noise model assumed, which is usually accomplished by calculating local moments over squared neighborhoods. When these neighborhoods are centered in pixels corresponding to image contours, the estimation is not accurate due to the presence of two or more tissues with different statistical properties. We overcome this limitation by introducing an anisotropic LMMSE scheme: the gray levels of each tissue in the MRI volume are modeled as a Gamma-mixture, such that we can discriminate between the different matters to construct anisotropic neighborhoods containing only one kind of tissue. The potential of the Gamma distribution relies on its ability to fit both the Rician distribution traditionally used to model the noise in MRI and the non-central Chi noise found in modern parallel MRI systems.

A realistic log-compressed law for ultrasound images based on a real device is proposed. The model takes into account the linear behavior of the logarithmic amplifier for small signal gain which transforms image values in a different way as the classical models do. Additionally, for recovery purposes, a method for the estimation of the compression parameters is also proposed when a realistic log-compressed law is considered. Results with synthetic images show that the proposed method achieved a consistent Rayleigh parameter estimate with a very low error. Experiments with real images show that the inversion method is consistent through the whole acquisition process when parameters of the logarithmic amplifier are assumed constant.

Diffusion Tensor Imaging (DTI) is an image modality that provides information about the nerve fiber tracts in the brain white matter. In this chapter, we describe the use of DTI to quantify the damage in fiber tracts due to brain tumors. A quantification method for this modality is useful to understand the neurological effects of the tumor in a given patient, such as motor and/or sensitive disorders, as well as to help in surgical planning for the resection of the tumor. The quantification of the state of fiber tracts and its comparison with the values obtained before resection and with healthy tracts, is also of paramount importance for the follow-up of the patient after tumor resection. As an example, we show the quantification of the pyramidal tract using two measures: integrity and connectivity. The fiber tract is automatically identified in any subject by means of a specific fiber tracking algorithm, so robust comparison between control subjects and patients can be performed, as well as comparison between the two hemispheres of a same subject. Diffusion tensor based measures are defined to quantify the state of fiber bundles, taking into account both the intrinsic properties of the fibers and the similarity to healthy control fibers. A 2D mapping of the tracts is also provided to perform comparisons among a given population, and to visually analyze the tract damages. Experiments show the results obtained in a set of 10 tumor patients, and 10 control subjects.

Classic geometric active contour algorithms have the limitation of segmenting the image into only two regions: background and object of interest. A new multiphase level set algorithm for the segmentation of two or more regions of interest is proposed. This algorithm avoids by construction the presence of overlapped and void regions and no additional coupling terms are required. In addition, the number of iterations needed to reach convergence is small. The algorithm has been tested against a state-of-the-art multiphase method on both simulated and real Magnetic Resonance Imaging (MRI) volumes with favorable results.

In this paper we generalize the Log-Euclidean polyaffine registration framework of Arsigny et al. to deal with articulated structures. This framework has very useful properties as it guarantees the invertibility of smooth geometric transformations. In articulated registration a skeleton model is defined for rigid structures such as bones. The final transformation is affine for the bones and elastic for other tissues in the image. We extend the Arsigny el al.'s method to deal with locally-affine registration of pairs of wires. This enables the possibility of using this registration framework to deal with articulated structures. In this context, the design of the weighting functions, which merge the affine transformations defined for each pair of wires, has a great impact not only on the final result of the registration algorithm, but also on the invertibility of the global elastic transformation. Several experiments, using both synthetic images and hand radiographs, are also presented.

In geometric active contour algorithms, a re-initialization step must be performed by the level set function to remain close to a signed distance function, in order to avoid numerical instabilities. We propose a new re-initialization method that may be employed as a standalone method to recover the signed distance condition, or may be embedded directly into a variational framework as an additional term for the energy functional. Its purpose is to make the pixels near the propagating contour be less affected by the re-initialization. Experimental results show that whereas previous approaches change the position of the zero level set, our method keeps it virtually unchanged.

A novel anisotropic diffusion filter is proposed in this work with application to cardiac ultrasonic images. It includes probabilistic models which describe the probability density function (PDF) of tissues and adapts the diffusion tensor to the image iteratively. For this purpose, a preliminary study is performed in order to select the probability models that best fit the stastitical behavior of each tissue class in cardiac ultrasonic images. Then, the parameters of the diffusion tensor are defined taking into account the statistical properties of the image at each voxel. When the structure tensor of the probability of belonging to each tissue is included in the diffusion tensor definition, a better boundaries estimates can be obtained instead of calculating directly the boundaries from the image. This is the main contribution of this work. Additionally, the proposed method follows the statistical properties of the image in each iteration. This is considered as a second contribution since state-of-the-art methods suppose that noise or statistical properties of the image do not change during the filter process.

In this paper we propose an alternative method to estimate and visualize the strain rate tensor (ST) in magnetic resonance images (MRI) when phase contrast MRI (PCMRI) and tagged MRI (TMRI) are not available. This alternative is based on image processing techniques. Concretely, an elastic image registration algorithm is used to estimate the movement of the myocardium at each point. Our experiments with real data prove that the registration algorithm provides a useful deformation field to estimate the ST fields. A classification between regional normal and dysfunctional contraction patterns, as compared with professional diagnosis, points out that the parameters extracted from the estimated ST can represent these patterns.

Image registration is a common image processing task, and therefore, many algorithms have been proposed and described to carry it out for different image modalities. However, the application of these algorithms to diffusion tensor imaging is not straightforward due to the special features of this kind of data, where a tensor is defined at each voxel. The information provided by the diffusion tensor is related to the anatomical structures in tissues, and this relation should be preserved, even though the image has been transformed by a registration procedure. On the other hand, the registration problem can be viewed as an optimization problem, where a similarity measure has to be maximized. The appropriate definition of this similarity measure is indeed an important issue for the registration of diffusion tensor images. In this paper, we compile the different approaches for the registration of diffusion tensor images that have been published. Special attention is paid to the aforementioned topics: how to preserve the coherence between the tensor and the underlying tissue structure, and how to measure the similarity between two diffusion tensors. Methods to evaluate results are also reviewed, since a reliable validation leads to more conclusive results, specially in the comparison of different techniques. Most challenging issues for diffusion tensor images registration are underlined, and open research lines about this topic are pointed out.

In this paper we deal with the problem of quantification of diffusion tensor (DT) data sets. A set of measures and a 2D tract mapping technique are proposed to analyze the fiber structures in brain white matter and to allow for comparisons between different subjects, either patients or controls. Features such as integrity, discontinuity and connectivity of the fiber bundles are proposed and analyzed, taking into account longitudinal and transverse information of the fiber bundle under study. The performance of the proposed characterization framework is shown analyzing the corticospinal tracts of control data sets and pathological cases, comparing the measures between controls and patients and also between the right and left hemispheres. A reproducibility study is also performed to show the robustness of the proposed measures.

Nowadays there is a growing interest in tensor medical imaging modalities. In Diffusion Tensor Magnetic Resonance Imaging (DT-MRI), each pixel is valued with a symmetric second-order tensor describing the spatial properties of diffusion at that point. Therefore, it provides significantly more information than scalar modalities, but this causes the complexity of the interfaces dealing with them to grow. In this chapter, the current situation of user interfaces for tensor fields is reviewed. Tensor user interfaces are difficult to design, given the difficulty of mentally integrating data with so many parameters. This is why a considerable effort must be invested in order to achieve intuitive and easy-to-use interfaces. The display of tensor information plays an important role in this, and we review several existing visualization methods for tensor fields.We must point out that, although most of the applications are graphical interfaces, there are also examples of command-line tools and multimodal interfaces employing virtual environments. We study some of the urrent medical user interfaces for diffusion tensor fields.

This chapter describes a framework for storage of tensor array data, useful to describe regularly sampled tensor fields. The main component of the framework, called Similar Tensor Array Core (STAC), is the result of a collaboration between research groups within the SIMILAR network of excellence. It aims to capture the essence of regularly sampled tensor fields using a minimal set of attributes and can therefore be used as a “greatest common divisor” and interface between tensor array processing algorithms. This is potentially useful in applied fields like medical image analysis, in particular in Diffusion Tensor MRI, where misinterpretation of tensor array data is a common source of errors. By promoting a strictly geometric perspective on tensor arrays, with a close resemblance to the terminology used in differential geometry, (STAC) removes ambiguities and guides the user to define all necessary information. In contrast to existing tensor array file formats, it is minimalistic and based on an intrinsic and geometric interpretation of the array itself, without references to other coordinate systems.

The evaluation of tensor image processing algorithms is an open problem that has not been broadly handled, and specific measures have not been described to assess the quality of tensor images. In this chapter, we propose the adaptation of quality measures that have been defined in the case of conventional scalar images to the tensor case, in order to evaluate the quality of the tensor images that are most frequently used in the image processing field. Special attention is paid to the tensor features that made this extension no straightforward. Some general concepts that should be taken into account for the definition of quality indexes for tensor images based on the well-known measures for conventional scalar images are detailed. Then, some of these measures are adapted to deal with tensor images and their behavior is analyzed by means of some examples. Thus, it is shown that structure based measures outperform point-wise measures, as well as the influence of handling all the tensor components.

In this paper we generalize the Log-Euclidean polyaffine registration framework of Arsigny et al. [1] to deal with articulated structures. This framework has very useful properties as it guarantees the invertibility of smooth geometric transformations. In articulated registration a skeleton model is defined for rigid structures such as bones. The final transformation is affine for the bones and elastic for other tissues in the image. We extend the Arsigny el al.'s method to deal with locally-affine registration of pairs of wires. This enables the possibility of using this registration framework to deal with articulated structures. In this context, the design of the weighting functions, which merge the affine transformations defined for each pair of wires, has a great impact not only on the final result of the registration algorithm, but also on the invertibility of the global elastic transformation. Several experiments, using both synthetic images and hand radiographs, are also presented.

A stochastic deformable model is proposed for the segmentation of the myocardium in Magnetic Resonance Imaging. The segmentation is posed as a probabilistic optimization problem in which the optimal time-dependent surface is obtained for the myocardium of the heart in a discrete space of locations built upon simple geometric assumptions. For this purpose, first, the left ventricle is detected by a set of image analysis tools gathered from the literature. Then, the segmentation solution is obtained by the Maximization of the Posterior Marginals for the myocardium location in a Markov Random Field framework which optimally integrates temporal-spatial smoothness with intensity and gradient related features in an unsupervised way by the Maximum Likelihood estimation of the parameters of the field. This scheme provides a flexible and robust segmentation method which has been able to generate results comparable to manually segmented images for some derived cardiac function parameters in a set of 43 patients affected in different degrees by an Acute Myocardial Infarction.

Since tensor usage has become more and more popular in image processing, the assessment of the quality between tensor images is necessary for the evaluation of the advanced processing algorithms that deal with this kind of data. In this paper, we expose the methodology that should be followed to extend well-known image quality measures to tensor data. Two of these measures based on structural comparison are adapted to tensor images and their performance is shown by a set of examples. By means of these experiments the advantages of structural based measures will be highlighted, as well as the need for considering all the tensor components in the quality assessment.

A method to identify different tissues in ultrasound images is presented. Assuming a Rayleigh model for speckle, the different tissues present in the image will be related to the Rayleigh sigma parameter, with different values for each different tissue. The parameter is locally estimated using well known estimation methods for the Rayleigh distribution, such as the maximum likelihood estimator. A significant increase in the separability of the different tissues present in the image is then achieved. A subsequent clustering process based on the position of the maxima of the estimator distribution allows the classification of each pixel in a tissue class.

In this paper, we present a set of techniques for the evaluation of brain tissue classifiers on a large data set of MR images of the head. Due to the difficulty of establishing a gold standard for this type of data, we focus our attention on methods which do not require a ground truth, but instead rely on a common agreement principle. Three different techniques are presented: the Williams’ index, a measure of common agreement; STAPLE, an Expectation Maximization algorithm which simultaneously estimates performance parameters and constructs an estimated reference standard; and Multidimensional Scaling, a visualization technique to explore similarity data. We apply these different evaluation methodologies to a set of eleven different segmentation algorithms on forty MR images. We then validate our evaluation pipeline by building a ground truth based on human expert tracings. The evaluations with and without a ground truth are compared. Our findings show that comparing classifiers without a gold standard can provide a lot of interesting information. In particular, outliers can be easily detected, strongly consistent or highly variable techniques can be readily discriminated, and the overall similarity between different techniques can be assessed. On the other hand, we also find that some information present in the expert segmentations is not captured by the automatic classifiers, suggesting that common agreement alone may not be sufficient for a precise performance evaluation of brain tissue classifiers.

Acquisition parameters play a crucial role in Diffusion Tensor Imaging (DTI), as they have a major impact on the values of scalar measures such as Fractional Anisotropy (FA) or Mean Diffusivity (MD) that are usually the focus of clinical studies based on white matter analysis. This paper presents an analysis on the impact of the variation of several acquisition parameters on these scalar measures with a novel double focus. First, a tractography-based approach is employed, motivated by the significant number of clinical studies that are carried out using this technique. Second, the consequences of simultaneous changes in multiple parameters are analyzed: number of gradient directions, b-value and voxel resolution. Results indicate that the FA is most affected by changes in the number of gradients and voxel resolution, while MD is specially influenced by variations in the b-value. Even if the choice of a tractography algorithm has an effect on the numerical values of the final scalar measures, the evolution of these measures when acquisition parameters are modified is parallel.

A segmentation application prototype of the vol- ume of the left ventricle for Magnetic Resonance Imaging is being developed. The foundation for this work is given by modeling possible radial deformations of the epicardium and endocardium contours by means of a Markov Random Field over which the most probable configuration is estimated. The field makes use of a Bayesian approach based on ap riori terms which impose smoothness along the coupled contours and likelihood terms which gather information provided by the images about the areas where the contours are supposed to be. The parameters of the field are estimated on a supervised basis. In this paper we propose a model-based segmentation procedure which is based on active contours and Markov Random Fields (MRF). Our solution is an adaptation of one previously developed for automatic extraction of kidney con- tours (6) in ultrasound images, which, in turn, was somehow inspired by the framework proposed in (7) for the 2D+T segmentation of endocardial and epicardial boundaries. In this paper, however we will analyze how a generic model to detect hard-to-tell boundaries (which is the case in ultrasound images) for star-like objects can be adapted to a different problem. Specifically, we focus on the problem of supervised parameter estimation from labeled data to determine whether the modeling terms are really needed or some of them may be removed or modified. Additionally, a new term (with respect to the model in (6)) is proposed to account for the joint presence of the two contours. Section II introduces the definitions needed in the paper and then both the prior and the likelihood functions of the model are described. Section III presents some quantitative results of the estimation of the model parameters and visual results of the segmentation of the contours. Section IV sum- marizes the paper together with some concluding remarks.

Practitioners' decision for mechanical aid discontinuation is a challenging task that involves a complete knowledge of a great number of clinical parameters, as well as its evolution in time. Recently, an increasing interest on respiratory pattern variability as an extubation readiness indicator has appeared. Reliable assessment of this variability involves a set of signal processing and pattern recognition techniques. This paper presents a suitability analysis of different methods used for breathing pattern complexity assessment. The contribution of this analysis is threefold: 1) to serve as a review of the state of the art on the so-called weaning problem from a signal processing point of view; 2) to provide insight into the applied processing techniques and how they fit into the problem; 3) to propose additional methods and further processing in order to improve breathing pattern regularity assessment and weaning readiness decision. Results on experimental data show that sample entropy outperforms other complexity assessment methods and that multidimensional classification does improve weaning prediction. However, the obtained performance may be objectionable for real clinical practice, a fact that paves the way for a multimodal signal processing framework, including additional high-quality signals and more reliable statistical methods.

We present three different sequential Wiener filters, namely, isotropic, orientation and anisotropic. The first one is similar to the classical Wiener filter in the sense that it uses an isotropic neighborhood to estimate its parameters. Here we present a sequential version of it. The orientation Wiener filter uses oriented neighborhoods to estimate the structure orientation present at each voxel, giving rise to a modified estimator of the parameters. Finally, the anisotropic Wiener filter combines both approaches adaptively so that the appropriate approach is locally selected. Several synthetic experiments are presented showing the performance of the filters with respect to their parameters. A mean square error analysis is performed using a publicly available magnetic resonance imaging (MRI) brain phantom and a comparison with other filtering approaches is carried out. In addition, results from filtering real MRI data are presented.

This chapter presents two techniques for regularization of tensor fields. We first present a nonlinear filtering technique based on normalized convolution, a general method for filtering missing and uncertain data. We describe how the signal certainty function can be constructed to depend on locally derived certainty information and further combined with a spatially dependent certainty field. This results in reduced mixing between regions of different signal characteristics, and increased robustness to outliers, compared to the standard approach of normalized convolution using only a spatial certainty field. We contrast this deterministic approach with a stochastic technique based on a multivariate Gaussian signal model in a Bayesian framework. This method uses a Markov random field approach with a 3D neighborhood system for modeling spatial interactions between the tensors locally. Experiments both on synthetic and real data are presented. The driving tensor application for this work throughout the chapter is the filtering of diffusion tensor MRI data.

In this paper, we present an evaluation of seven automatic brain tissue classifiers based on level of agreements. A number of agreement measures are explained, and we show how they can be used to compare different segmentation techniques. We use the Simultaneous Truth and Performance Level Estimation (STAPLE) of Warfield et al. but also introduce a novel evaluation technique based on the Williams' index. The methods are evaluated using these two techniques on a population of forty subjects, each having an SPGR scan and a co-registered T2 weighted scan. We provide an interpretation of the results and show how similar the output of the STAPLE analysis and Williams' index are. When no ground truth is required, we recommend the use of Williams' index as it is easy and fast to compute.

This paper presents an algorithm for automatically detecting bone contours from hand radiographs using active contours. Prior knowledge is first used to locate initial contours for the snakes inside each bone of interest. Next, an adaptive snake algorithm is applied so that parameters are properly adjusted for each bone specifically. We introduce a novel truncation technique to prevent the external forces of the snake from pulling the contour outside the bones boundaries, yielding excellent results.

Fractional low order moments have been reported as beneficial for sampling computations using the K distribution. However, it has been recently pointed out that this it not the case for the homodyned-K distribution for a tissue discrimination problem. In this paper we show that such an statement is not fully justified. To that end, we follow a standard pattern recognition procedure both to determine class separability measures and to classify data with several classifiers. We conclude that the optimum order of the moments is intimately linked to the specific statistical properties of the tissues to be discriminated. Some ideas on how to choose the optimum order are discussed.

In this paper we propose a novel lossless coding scheme for medical images that allows the final user to switch between a lossy and a lossless mode. This is done by means of a progressive reconstruction philosophy (which can be interrupted at will) so we believe that our scheme gives a way to trade off between the accuracy needed for medical diagnosis and the information reduction needed for storage and transmission. We combine vector quantization, run-length bit plane and entropy coding. Specifically, the first step is a vector quantization procedure; the centroid codes are Huffman- coded making use of a set of probabilities that are calculated in the learning phase. The image is reconstructed at the coder in order to obtain the error image; this second image is divided in bit planes, which are then run-length and Huffman coded. A second statistical analysis is performed during the learning phase to obtain the parameters needed in this final stage. Our coder is currently trained for hand-radiographs and fetal echographies. We compare our results for this two types of images to classical results on bit plane coding and the JPEG standard. Our coder turns out to outperform both of them.© (2000) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

It has been shown that the tensor calculation is very sensitive to the presence of noise in the acquired images, yielding to very low-quality Diffusion Tensor Images (DTI) data. Recent investigations have shown that the noise present in the Diffusion Weighted Images (DWI) causes bias effects on the DTI data which cannot be corrected if the noise characteristic is not taken into account. One possible solution is to increase the minimum number of acquired measurements (which is 7) to several tens (or even several hundreds). This has the disadvantage of increasing the acquisition time by one (or two) orders of magnitude, making the process inconvenient for a clinical setting. We here proposed a turn-around procedure for which the number of acquisitions is maintained but, the DWI data are filtered prior to determining the DTI. We show a significant reduction on the DTI bias by means of a simple and fast procedure which is based on linear filtering; well-known drawbacks of such filters are circumvented by means of anisotropic neighborhoods and sequential application of the filter itself. Information of the first order probability density function of the raw data, namely, the Rice distribution, is also included. Results are shown both for synthetic and real datasets. Some error measurements are determined in the synthetic experiments, showing how the proposed scheme is able to reduce them. It is worth noting a 50% increase in the linear component for real DTI data, meaning that the bias in the DTI is considerably reduced. A novel fiber smoothness measure is defined to evaluate the resulting tractography for real DWI data. Our findings show that after filtering, fibers are considerably smoother on the average. Execution times are very low as compared to other reported approaches which allows for a real-time implementation.

3D freehand ultrasound imaging is becoming a widespread technique in medical examinations. This imaging technique produces a set of irregularly spaced B-scans. Reconstructing a regular grid from these B-scans is a challenging problem that enables the visualization and further analysis of the acquired data. This paper focuses on extending an existing method [1] to define the output reconstruction grid based on principal component analysis (PCA). Our method introduces a model for the region of interest (ROI) in order to adapt the grid to the ROI. In addition, a technique based on normalized convolution is proposed for the interpolation problem. A new applicability function based on the correlation function of a linear probe is used to avoid inter-resolution cell blurring.

In this paper we propose a novel non-Gaussian MRF for regularization of tensor fields for fiber tract enhancement. Two entities are considered in the model, namely, the linear component of the tensor, i.e., how much line-like the tensor is, and the angle of the eigenvector associated to the largest eigenvalue. A novel, to the best of the author’s knowledge, angular density function has been proposed. Closed form expressions of the posterior densities are obtained. Some experiments are also presented for which color-coded images are visually meaningful. Finally, a quantitative measure of regularization is also calculated to validate the achieved results based on an averaged measure of entropy.