Sylvie Sevestre-Ghalila

Nakagami Markov random field as texture model for ultrasound RF envelope image

The aim of this paper is to propose a new Markov random field (MRF) model for the backscattered ultrasonic echo in order to get information about backscatter characteristics, such as the scatterer density, amplitude and spacing. The model combines the Nakagami distribution that describes the envelope of backscattered echo with spatial interaction using MRF. In this paper, the parameters of the model and the estimation parameter method are introduced. Computer simulation using ultrasound radio-frequency (RF) simulator and experiments on choroidal malignant melanoma have been undertaken to test the validity of the model. The relationship between the parameters of MRF model and the backscatter characteristics has been established. Furthermore, the ability of the model to distinguish between normal and abnormal tissue has been proved. All the results can show the success of the model.

Texture Image Analysis for Osteoporosis Detection with Morphological Tools

Osteoporosis is due to the following two phenomena: a reduction bone mass and a degradation of the microarchitecture of bone tissue. In this paper, we propose a method for extracting morphological information enabling the description of bone structure from radiological images of the calcaneus. Our main contribution relies on the fact that we provide bone descriptors close to classical 3D-morphological bone parameters. The first step of the proposed method consists in extracting the grey-scale skeleton of the microstructures contained in the underlying images. After an appropriate processing, the resulting skeleton provides discriminant features between osteoporotic patients and control patients. Statistical tests corroborate this discrimination property.

Ultrasound Backscatter Characterization by Using Markov Random Field Model

This paper evaluates a K-Markov random field model for retrieving information about backscatter characteristics, especially regularity spacing scatterers in simulated ultrasound image. The model combines a statistical K-distribution that describes the envelope of backscattered echo and spatial interaction given by Markov random field (MRF). Parameters estimated by the conditional least squares (CLS) estimation method on simulated radio-frequency (RF) envelope image show that the interaction parameters measure the degree of the randomness of the scatterers

Graph Structuring of Skeleton Object for Its High-Level Exploitation

Skeletonization is a morphological operation that summarizes an object by its median lines while preserving the initial image topology. It provides features used in biometric for the matching process, as well as medical imaging for quantification of the bone microarchitecture. We develop a solution for the extraction of structural and morphometric features useful in biometric, character recognition and medical imaging. It aims at storing object descriptors in a re-usable and hierarchical format. We propose graph data structures to identify skeleton nodes and branches, link them and store their corresponding features. This graph structure allows us to generate CSV files for high level analysis and to propose a pruning method that removes spurious branches regarding their length and mean gray level. We illustrate manipulations of the skeleton graph structure on medical image dedicated to bone microarchitecture characterization.

AffectiveROAD system and database to assess driver's attention

Thanks to the rise of new wearable and non-intrusive sensor technology, Internet of Things (IoT) contributes in human daily life improvement. In the context of smart vehicles, human affective monitoring should be based on a context-aware system in order to consider the interactions between the driver, the vehicle and the ambient environment. In this paper, we propose AffectiveROAD platform, that senses the human physiological changes, the ambient environment inside the vehicle, and the vehicle speed. Thanks to this platform, several drivers state indicators such as stress and arousal may be developed and validated. Two types of wireless physiological sensors are used to monitor the electrodermal activity, the heart rate, the skin temperature, the respiration, and the hand movement of the driver. Moreover, we developed a sensor network allowing to capture the ambient temperature, humidity, pressure, and luminosity. The purpose of this paper is to describe a real-world driving protocol allowing to collect data using IoT-based materials and to announce the publication of a database for drivers state monitoring research. A partial database concerning the physiological and the environmental information is available on request, for public use.

The Benefit of a Kernel Estimate Based Forward Projection for Iterative Tomographic Reconstruction Techniques

X-ray forward and backward projections are major steps in all iterative tomographic reconstruction techniques. To ensure the quality of the reconstruction results, the projector should fit the imaging system geometry and the x-ray physic as well as possible. In this paper, we intend to show the benefit of our proposed kernel estimate based forward projection for iterative tomographic reconstruction techniques, such as Algebraic Reconstruction Techniques (ART). The proposed projection method is independent to the volume sampling and takes into account the projection geometry (cone-beam and parallel) by simply adjusting the variable bandwidth of the kernel estimate. We show that the sample-point estimate (adaptive-bandwidth kernel estimate) based forward projection provides accurate projections and better reconstructions than the fixed bandwidth estimate based forward projection which is similar to the well known Splatting techniques.

Random Forest-Based Approach for Physiological Functional Variable Selection: Towards Driver's Stress Level Classification

This paper deals with physiological functional variables selection for driver’s stress level classification using random forests. Our analysis is performed on experimental data extracted from the drivedb open database available on PhysioNet website. The physiological measurements of interest are: electrodermal activity captured on the driver’s left hand and foot, electromyogram, respiration, and heart rate, collected from ten driving experiments carried out in three types of routes (rest area, city, and highway). The contributions of this work touch on the method as well as the application aspects. From a methodological viewpoint, the physiological signals are considered as functional variables, decomposed on a wavelet basis and then analyzed in search of most relevant variables. On the application side, the proposed approach provides a “blind” procedure for driver’s stress level classification, giving close performances to those resulting from the expert-based approach, when applied to the drivedb database. It also suggests new physiological features based on the wavelet levels corresponding to the functional variables wavelet decomposition. Finally, the proposed approach provides a ranking of physiological variables according to their importance in stress level classification. For the case under study, results suggest that the electromyogram and the heart rate signals are less relevant compared to the electrodermal and the respiration signals. Furthermore, the electrodermal activity measured on the driver’s foot was found more relevant than the one captured on the hand. Finally, the proposed approach also provided an order of relevance of the wavelet features.

SHAPE AND REGULARITY OF 3D CORTICAL BONE CANALS: COMPARAISON BETWEEN DESKTOP AND SYNCHROTRON RADIATION MICRO-CT IMAGES

The osteoporotic fractures are due to alteration of trabecular and/or cortical bone. As the trabecular bone analysis do not require images with high resolution, it has been largely investigated contrary to the cortical bone for which there are few studies. The porosity of the cortical bone is depending on canal network and the bone remodeling leads to a coalescence process of canals which progressively involve the degradation of bone quality. Most of parameters to characterize the canal network of cortical bone were derived from the trabecular bone micro-architecture parameters and do not characterize the canals locally for which the changes of shape and regularity are in relation to the remodelling intensity. In this preliminary study, new tri-dimensional (3D) features of shape and surface at the canal level are presented to better characterize the remodeling processes. For that, we used 16 cortical bone samples from human femurs. Three D images at 7.5 μm of resolution were obtained by synchrotron radiation micro-CT (SR micro-CT) used as the reference method comparatively to desktop micro-CT usually used for bone exploration. Therefore, our goal is to compare these both modalities. Usual and global parameters about canal volume, surface, diameter, spacing and number derived from the trabecular bone analysis were performed. Two parameters were developed to better describe the complexity of the structure: one describing the shape of the canals Ca.s and one the regularity surface Ca.r. There are significant changes in shape and surface parameters on D micro-CT compared with those observed by the reference SR micro-CT whereas there was no difference for global parameters. Consequently, only high image quality, such as SR micro-CT modality could be used to characterize the cortical degradation process at a local level.

Quantification of microarchitecture bone junctions based on a dual tree M-band wavelet decomposition

Osteoporosis shows itself both in a reduction of the bone mass and a degradation of the microarchitecture of the bone tissue. To this respect, radiographies of the calcaneus are used to analyze both the texture and the structure of the bone thanks to sophisticated image processing tools. In this paper, we propose a method for evaluating the number of junctions in the imaged microarchitecture. The first novelty of this paper is the evaluation of this number from a multiresolution representation resulting from Dual Tree M-band decompositions. Its appealing advantage is its great directional selectivity. The second contribution of our work relies on the statistical procedure we apply to separate between osteoporotic patients (OP) and control patients (CP). Classification and statistical tests conducted on a set of radiographies with their own ground truth corroborate the advantage of the proposed method.