Time filter

Source Type

Roujol S.,University of Bordeaux Segalen | Roujol S.,CNRS Computer Science Laboratory | De Senneville B.D.,University of Bordeaux Segalen | Hey S.,University of Bordeaux Segalen | And 2 more authors.
IEEE Transactions on Medical Imaging | Year: 2012

Real time magnetic resonance (MR) thermometry is gaining clinical importance for monitoring and guiding high intensity focused ultrasound (HIFU) ablations of tumorous tissue. The temperature information can be employed to adjust the position and the power of the HIFU system in real time and to determine the therapy endpoint. The requirement to resolve both physiological motion of mobile organs and the rapid temperature variations induced by state-of-the-art high-power HIFU systems require fast MRI-acquisition schemes, which are generally hampered by low signal-to-noise ratios (SNRs). This directly limits the precision of real time MR-thermometry and thus in many cases the feasibility of sophisticated control algorithms. To overcome these limitations, temporal filtering of the temperature has been suggested in the past, which has generally an adverse impact on the accuracy and latency of the filtered data. Here, we propose a novel filter that aims to improve the precision of MR-thermometry while monitoring and adapting its impact on the accuracy. For this, an adaptive extended Kalman filter using a model describing the heat transfer for acoustic heating in biological tissues was employed together with an additional outlier rejection to address the problem of sparse artifacted temperature points. The filter was compared to an efficient matched FIR filter and outperformed the latter in all tested cases. The filter was first evaluated on simulated data and provided in the worst case (with an approximate configuration of the model) a substantial improvement of the accuracy by a factor 3 and 15 during heat up and cool down periods, respectively. The robustness of the filter was then evaluated during HIFU experiments on a phantom and in vivo in porcine kidney. The presence of strong temperature artifacts did not affect the thermal dose measurement using our filter whereas a high measurement variation of 70% was observed with the FIR filter. © 2011 IEEE.

Guizard N.,Montreal Neurological Institute | Coupe P.,CNRS Computer Science Laboratory | Fonov V.S.,Montreal Neurological Institute | Manjon J.V.,Polytechnic University of Valencia | And 2 more authors.
NeuroImage: Clinical | Year: 2015

Multiple sclerosis (MS) lesion segmentation is crucial for evaluating disease burden, determining disease progression and measuring the impact of new clinical treatments. MS lesions can vary in size, location and intensity, making automatic segmentation challenging. In this paper, we propose a new supervised method to segment MS lesions from 3D magnetic resonance (MR) images using non-local means (NLM). The method uses a multi-channel and rotation-invariant distance measure to account for the diversity of MS lesions. The proposed segmentation method, rotation-invariant multi-contrast non-local means segmentation (RMNMS), captures the MS lesion spatial distribution and can accurately and robustly identify lesions regardless of their orientation, shape or size. An internal validation on a large clinical magnetic resonance imaging (MRI) dataset of MS patients demonstrated a good similarity measure result (Dice similarity = 60.1% and sensitivity = 75.4%), a strong correlation between expert and automatic lesion load volumes (R2 = 0.91), and a strong ability to detect lesions of different sizes and in varying spatial locations (lesion detection rate = 79.8%). On the independent MS Grand Challenge (MSGC) dataset validation, our method provided competitive results with state-of-the-art supervised and unsupervised methods. Qualitative visual and quantitative voxel- and lesion-wise evaluations demonstrated the accuracy of RMNMS method. © 2015 The Authors. Published by Elsevier Inc.

Boubchir L.,Qatar University | Boubchir L.,CNRS Computer Science Laboratory | Boashash B.,Qatar University | Boashash B.,University of Queensland
IEEE Transactions on Signal Processing | Year: 2013

This paper presents a novel nonparametric Bayesian estimator for signal and image denoising in the wavelet domain. This approach uses a prior model of the wavelet coefficients designed to capture the sparseness of the wavelet expansion. A new family of Bessel K Form (BKF) densities are designed to fit the observed histograms, so as to provide a probabilistic model for the marginal densities of the wavelet coefficients. This paper first shows how the BKF prior can characterize images belonging to Besov spaces. Then, a new hyper-parameters estimator based on EM algorithm is designed to estimate the parameters of the BKF density; and, it is compared with a cumulants-based estimator. Exploiting this prior model, another novel contribution is to design a Bayesian denoiser based on the Maximum A Posteriori (MAP) estimation under the 0-1 loss function, for which we formally establish the mathematical properties and derive a closed-form expression. Finally, a comparative study on a digitized database of natural images and biomedical signals shows the effectiveness of this new Bayesian denoiser compared to other classical and Bayesian denoising approaches. Results on biomedical data illustrate the method in the temporal as well as the time-frequency domain. © 1991-2012 IEEE.

Eskildsen S.F.,University of Aarhus | Coupe P.,CNRS Computer Science Laboratory | Fonov V.S.,Montreal Neurological Institute | Pruessner J.C.,McGill University | Collins D.L.,Montreal Neurological Institute
Neurobiology of Aging | Year: 2015

Optimized magnetic resonance imaging (MRI)-based biomarkers of Alzheimer's disease (AD) may allow earlier detection and refined prediction of the disease. In addition, they could serve as valuable tools when designing therapeutic studies of individuals at risk of AD. In this study, we combine (1) a novel method for grading medial temporal lobe structures with (2) robust cortical thickness measurements to predict AD among subjects with mild cognitive impairment (MCI) from a single T1-weighted MRI scan. Using AD and cognitively normal individuals, we generate a set of features potentially discriminating between MCI subjects who convert to AD and those who remain stable over a period of 3 years. Using mutual information-based feature selection, we identify 5 key features optimizing the classification of MCI converters. These features are the left and right hippocampi gradings and cortical thicknesses of the left precuneus, left superior temporal sulcus, and right anterior part of the parahippocampal gyrus. We show that these features are highly stable in cross-validation and enable a prediction accuracy of 72% using a simple linear discriminant classifier, the highest prediction accuracy obtained on the baseline Alzheimer's Disease Neuroimaging Initiative first phase cohort to date. The proposed structural features are consistent with Braak stages and previously reported atrophic patterns in AD and are easy to transfer to new cohorts and to clinical practice. © 2015 Elsevier Inc.

Poizat P.,University of Evry Val dEssonne | Poizat P.,CNRS Computer Science Laboratory | Salaun G.,Grenoble Institute of Technology
Proceedings of the ACM Symposium on Applied Computing | Year: 2012

Choreographies allow business and service architects to specify with a global perspective the requirements of applications built over distributed and interacting software entities. While being a standard for the abstract specification of business workflows and collaboration between services, the Business Process Modeling Notation (BPMN) has only been recently extended into BPMN 2.0 to support an interaction model of choreography, which, as opposed to interconnected interface models, is better suited to top-down development processes. An important issue with choreographies is real-izability, i.e., whether peers obtained via projection from a choreography interact as prescribed in the choreography requirements. In this work, we propose a realizability checking approach for BPMN 2.0 choreographies. Our approach is formally grounded on a model transformation into the LOTOS NT process algebra and the use of equivalence checking. It is also completely tool-supported through interaction with the Eclipse BPMN 2.0 editor and the CADP process algebraic toolbox. © 2012 ACM.

Discover hidden collaborations