Jean-Yves Boire

A General Framework for the Fusion of Anatomical and Functional Medical Images

The collection of various data coming from anatomical and functional imagery is becoming very common for the study of a given pathology, and their aggregation generally allows for a better medical decision in clinical studies. A fusion process is described in this article for the modeling of this aggregation. The process is illustrated in the case of anatomical and functional images of the brain, but the general principle may be extended to other organs. The whole three-step fusion process based on possibilistic logic is detailed and a new class of fusion operator is introduced. The use of fuzziness in the process in general and in the operator in particular allows for the management of uncertainty and imprecision inherent to the images. The fusion process is illustrated in two clinical cases: the study of Alzheimers disease by MR/SPECT fusion and the study of epilepsy by MR/SPECT/PET fusion. Results are presented and evaluated, and a preliminary clinical validation is achieved. The assessment of the method is encouraging, allowing its application on several clinical problems.

Tissue segmentation on MR images of the brain by possibilistic clustering on a 3D wavelet representation

An algorithm for the segmentation of a single sequence of three‐dimensional magnetic resonance (MR) images into cerebrospinal fluid, gray matter, and white matter classes is proposed. This new method is a possibilistic clustering algorithm using the fuzzy theory as frame and the wavelet coefficients of the voxels as features to be clustered. Fuzzy logic models the uncertainty and imprecision inherent in MR images of the brain, while the wavelet representation allows for both spatial and textural information. The procedure is fast, unsupervised, and totally independent of any statistical assumptions. The method is tested on a phantom image, then applied to normal and Alzheimers brains, and finally compared with another classic brain tissue segmentation method, affording a relevant classification of voxels into the different tissue classes. J. Magn. Reson. Imaging 2000;11:267–278.

Segmentation of fat and muscle from MR images of the thigh by a possibilistic clustering algorithm.

Physical training is proved to induce changes in physical capacity and body composition. We propose in this article a fast, unsupervised and fully three-dimensional automatic method to extract muscle and fat volumes from magnetic resonance images of thighs in order to assess these changes. The technique relies on the use of a fuzzy clustering algorithm and post-processings to accurately process the body composition of thighs. Results are compared on 11 healthy voluntary elderly people with those provided on the same data by a validated method already published, and its reliability is assessed on repeated measures on three subjects. The two methods statistically agree when computing muscle and fat volumes, and clinical implications of this fully automatic method are important for medicine, physical conditioning, weight-loss programs and predictions of optimal body weight.

MRI–SPECT fusion for the synthesis of high resolution 3D functional brain images: a preliminary study

Medical imaging being a fast-expanding field, multimodal data fusion appears more and more as a key element for the optimal use of images. By fusion, we mean the combination of several information sources (in particular images), with the aim of providing either more condensed or more pertinent information. The long term scope of this work would be to improve the interpretation of 3D brain images, providing extra elements for the diagnosis and patient follow up. This preliminary study is part of a wider context: the medical follow up of patients suffering from probable Alzheimer disease observed in single photon emission tomography by fusion after registration with magnetic resonance images. Several information combination techniques based on the possibility theory are presented. A new operator, more specifically adapted to the fusion of anatomical and functional images, as well as a high resolution functional image synthesis technique are proposed. A first comparative study of fusion techniques is then proposed. Although no thorough test protocol has been defined, these preliminary results are encouraging, giving access to a wide field of potential clinical applications.

Automatic volumetric measurement of lateral ventricles on magnetic resonance images with correction of partial volume effects

To propose a method for the quantification of lateral ventricle (LV) volumes on a single sequence of 3D magnetic resonance (MR) images.

Automatic assessment of muscle/fat temporal variations on MR images of the thigh

An automatic segmentation method for the quantification of muscle, subcutaneous and intermuscular fat from MR images of thighs is described. This process is applied to the study of thigh physiology evolution during nutritional follow-up and physical training.

Data fusion and fuzzy spatial relationships for locating deep brain stimulation targets in magnetic resonance images

Symptoms of Parkinsons disease can be relieved through Deep Brain Stimulation. This neurosurgical technique relies on high precision positioning of electrodes in specific areas of the basal ganglia and the thalamus. In order to identify these anatomical targets, which are located deep within the brain, we developed a semi-automated method of image analysis, based on data fusion. Information provided by both anatomical magnetic resonance images and expert knowledge is managed in a common possibilistic frame, using a fuzzy logic approach. More specifically, a graph-based virtual atlas modeling theoretical anatomical knowledge is matched to the image data from each patient, through a research algorithm (or strategy) which simultaneously computes an estimation of the location of every structures, thus assisting the neurosurgeon in defining the optimal target. The method was tested on 10 images, with promising results. Location and segmentation results were statistically assessed, opening perspectives for enhancements.

Monitoring and modeling of pupillary dynamics

The aim of the current study is the conception of a tool to analyze the pupillary dynamics. The reading of the pupil motion is a well-known method that provides functional information about the autonomous nervous system. The monitoring is made up of three parts. First, a digital camera is chosen for the image acquisition because of its simple use at the patients bed and its low cost. Then a software estimates the pupillogram, i.e. the plot of the pupil area versus time, from the digital video sequence using deformable templates. And finally, the pupil light reflex (PLR) is modeled by a second order linear system to compute characteristic parameters. They can be used to diagnose various affections from a statistical analysis.

Synthesis of a Functional Information with Anatomical Landmarks by Multiresolution Fusion of Brain Images

In order to help clinicians with the diagnosis of neurodegenerative diseases, we provide a synthetic functional information located in relation with anatomical structures. The final image is processed by multimodal data fusion between SPECT and MR images. We propose a new method for the management of such multiresolution data, in which a geometrical model allows an accurate correspondence of voxels from both images, while preserving at best both original pieces of information. We use this matching method to replace the interpolation step in the compulsory image registration of the data fusion process. The geometrical model is first built from registration parameters. Computational geometry algorithms, applied to this model, allow the computation of numerical values used to process the final information. The method has been applied to brain perfusion and neurotransmission SPECT images

Segmentation of the striatum using data fusion

Proposes a new segmentation scheme to detect cerebral structures in MRI acquisitions using numerical information contained in the image and expert knowledge brought by a specialist. This process is divided in three steps: first, information contained in the MR image is extracted using a fuzzy clustering algorithm, and theoretical information concerning the structure to segment is modeled using possibility theory. Information fusion is then processed, followed by a decision step ending the structure segmentation. Heads of caudate nuclei and putamens are segmented using this method. Results are promising and validation is performed using both numerical indexes and assessment by an expert. This method can be applied to any cerebral structure in an MR image, provided that it can be described in terms of shape, direction and distance by an expert and that the contrast and resolution of the MRI are sufficient.