Search

In the era of the Internet of Things (IoT) and Big Data (BD), a significant amount of data is permanently generated every day. The data size of collected data streams is now reaching zetta bytes (i.e., 1021), and their processing and analysis becomes more and more challenging especially in embedded systems, where the overall goal is to maximize performance per watt, while meeting real-time requirements and trying to keep the overall power consumption in the very limited power budgets. The collected data are often reduced by means of clustering, vector quantization or compression before their further processing. The unsupervised learning techniques such as Self-Organizing Maps (SOMs) not needing any prior knowledge of processed data are perfect candidates for this task. However, real-time vector quantization with SOMs requires high performances and dynamic online configurability. The software counterparts of SOMs are highly flexible with limited performances per watt whereas the hardware SOMs generally lack of flexibility. In this paper, a novel scalable, dynamic and growing hardware self-organizing map (SOM) is presented. The presented hardware SOM architecture is dynamically configurable and adaptable in terms of neurons, map size and vector dimension depending on the application-specific needs. The proposed architecture is validated on different map sizes (up to 16×16) with different vector widths applied for real-time color quantization and pattern distribution recognition.

The LoRaWAN technology is today the object of great interest in the Internet of Things context. It defines a simple network architecture offering a wide-area wireless coverage for low rate IoT applications with low power consumption for devices. The LoRaWan class A is designed for sensor networks with a focus on the uplink. LoRaWan defines an optional MAC operation, Class B, that provides the network server with the opportunities to initiate a downlink, which can be a real solution for actuators focus network. Today, Performances of Class B are not quantified and compared to default LoRaWAN class. In this paper, we propose an evaluation of Class B performance. We offer a set of realistic evaluation scenarios based on an NS-3 simulation module that we have developed for this purpose. Results show that Class B reduces the delivery delay of downlink traffic in comparison to Class A. Class B operation significantly reduces the percentage of packet loss for downlink traffic even in congested contexts. We conclude that a trade off should be made between having low access delay or packet loss. Both the NS-3 module and data are released as an open-source to the research community.

In this paper, electrocardiogram (ECG) analyses were used as valuable a tool in the evaluation of cognitive tasks also given by the electroencephalograms (EEG). By taking and analyzing measurements in large quantities, we tried to better understand the functioning of human physiological systems. This study examined the cognitive and cardiovascular system function simultaneously. The purpose of this paper was to seek statistical causality in the sense of Granger between the EEG and ECG signals based on time series and autoregressive vector processes (VAR). For this purpose, 24 hours were recorded and during the tests, random and non-stationary portions of the ECG and EEG were extracted. The results indicated that there was Granger causality between the signals. This allowed us to forecast and predict traffic spots within and between the ECG and EEG signals.

In this paper we present a new architectural approach of a Self-Organizing Map (SOM). The proposed architecture, called Systolic-SOM (SSOM), is based on a generic formalism that exploits two levels of nested parallelism of neurons and connections. Thus, this solution provides a distributed set of independent computations between the neuroprocessors that define the SSOM architecture. To validate our approach, we evaluate the performance of several SOM network architectures after their integration on FPGA support. This architecture has achieved a performance almost twice as fast as that obtained in recent literature.

In this work, a new method for removal of ocular and muscular artifacts from Multi-channel electroencephalogram (EEG) is presented in order to obtain a 3D filtered cerebral mapping images. First, a FastICA algorithm of Independent Component Analysis (ICA) is applied in combination with the tunable Q-factor wavelet transform (TQWT), FastICA-TQWT. Then, to show the robustness of this method, a comparison was made between the proposed FastICA-TQWT method and the classical one FastICA-DWT based on three criteria: Mean Squared Error (MSE), correlation coefficient and Signal Noise Ratio (SNR). The results showed that the FastICA-TQWT method gave the highest Signal Noise Ratio and correlation coefficient and the minimum Mean Squared Error. However, the FastICA-TQWT algorithm requires an extremely high computing power. Therefore, the second contribution of this paper is to provide an EEG signal treatment by implementing the hybrid FastICA-TQWT algorithm using a new computing technology designed for a high-performance computing, called Graphical Processing Units (GPUs) using the Compute Unified Device Architecture (CUDA) technology. The performance of the parallel approach running along the GPU was compared to a CPU implementation.

Heart rates have normal values ranging from 60 to 80 beats per minute (bpm) for adults. RR distances have normal durations between 0.75 and 1 second. The complexes QRS durations have normal durations of less than 0.1 second. However, heart rate and RR distances also depend on age (adult or child), the patient's status (rest or stress), sex (male or female) and the patient's conditions (healthy or pathological). Heart rates, RR distances and QRS durations are not sufficient to determine whether ECGs are normal or pathological. Recently, two novel metrics have been calculated to reflect the regularity of RR distances and the QRS complexes durations irrespective of the patient's age, sex and heart rate. In this paper, these novel parameters were tested and validated on the arrhythmia (MIT-BIH), Abdominal and Direct Fetal ECG (ADFECGDB) and BIDMC Congestive Heart Failure (CHFDB) databases.

Conformable electronics, electronics that can be integrated into ergonomically or aesthetically 2.5/3D shape surfaces, have received a tremendous attention during recent years. So far, various materials, designs and fabrication methods were developed for the realization of dynamically-stretchable (elastic) circuits.

The choice of a segmentation method depends on several considerations, namely the nature of the image, the primitives to extract and the segmentation methods. We propose an MLP-basis neuronal approach for the choice of the segmentation method taking into account the nature of the input image. First, an evaluation of the quality of segmentation by different methods and using various criteria of evaluation was carried out. Then, a characterization of images, based on some objective parameters, was performed. The resulting descriptors will be used as input to the neuronal approach to associate each type of image with the adequate segmentation method after learning. We report the results of the intelligent segmentation method choice obtained on different databases of medical images. The discussion of these encouraging results allowed us to improve our success rate and cover all varieties of images.

Today a need is emerging for embedding electronic and sensor functions in the products which needs these functions, and, importantly, to do this without noticeably influencing the mechanical design of the product. This contribution describes an approach used to produce a 2.5/3D free-form rigid and smart objects or randomly shaped circuit. The proposed fabrication process of shaped circuit is compatible with a typical printed circuit manufacturing and electronics assembly. Once the circuit is completed in its flat shape, its random final functional shape is given using thermoforming. In order to be able to deform a given flat circuit to its final form with predictable final spatial positions of components and interconnections. A FEM simulation is conducted to model the thermoforming of polymer based electronic circuits. As one of the process outputs, the wall thickness distribution predicted for the final part is compared with the experimental results.

This paper deals with patient-specific blood flow modeling in a stenosed internal carotid artery (ICA). An ICA stenosis has an impact on hemodynamic behavior. It can hamper the brain irrigation and even cause a stroke. Our aim is to predict the blood flow behavior through computational fluid dynamic (CFD) study. The proposed approach realizes a hemodynamic modeling within a geometric carotid model build from a 3D computed tomography angiography image with blood considered as a Newtonian and incompressible fluid, and the wall as rigid. The blood flow modeling is based on the Navier-Stokes equation. A Womersley velocity profile is used as a boundary condition in the common carotid artery. Main results of this study are the following: (1) velocity is maximum in stenosis and minimum in the sinus, (2) pressure is negative on the most of carotid artery bifurcation unless the post-stenotic site.

Osteoporosis is a disease characterized by low bone mass and deterioration of micro-architectural bone tissue, which provokes an increased risk of fracture. This work treats the texture characterization of trabecular bone radiographs. The goal is to analyse according to clinical research a group of 174 subjects: 87 osteoporotic patients with various bone fracture types and 87 healthy subjects. In order to characterize osteoporosis, a method of lacunarity measurement for grayscale image is used. This approach allowed the discrimination between healthy subjects and patients with osteoporosis. The results show an improved classification rate compared to another work [1].

This paper presents a robust method of local deformation analysis of the heart's left ventricle (LV) aimed at specifying the affected area. Our method is based on regional curvature variation calculation using the Hotelling T2 two samples difference metric. It has been validated with real data obtained from myocardial scintigraphy imaging techniques of 44 patients. For each patient we extract 3D surfaces representing the anatomical structures of interest, namely the endocardium and epicardium at stress. In order to progress to a regional analysis, we carried out, for each patient, a division into 17 regions according to the AHA standard. Experimental results demonstrate the great robustness and efficiency of our method.

The aim of this study is the detection and segmentation of microcalcifications in digital mammograms using multifractal analysis. To detect the suspicious Region Of Interest (ROI), containing anomalies, we propose to decompose the whole image into ROIs and compare the multifractal spectrums based on the q-structure functions of each one. The segmentation of microcalcifications consists of two steps. On the first step, we create an image denoted ‘α_image’. This image is constructed using the singularity coefficient, deduced from multifractal spectrum. Then, in the next step, we enhance the visualization of microcalcifications by creating an image denoted ‘f(α)_image’ based on the global regularity measure of the ‘α_image’ spectrum. We investigated the robustness of our approach using a data set of mammograms from ‘MiniMIAS’ database. Results demonstrate the accuracy of our approach, which successfully detect and segment microcalcifications with irregular form and small size.

A closed surface of a 3D object with spherical topology can be expressed into a series of spherical harmonic (SPHARM) functions. The SPHARM descriptors allow accurate representation of large data sets with small number of coefficients. In this paper, we present a new approach for regional deformation analysis of the heart left ventricle (LV). We use spherical harmonics (SPHARM) shape descriptors with the modified Hotelling T2 metric in order to characterize the myocardium disease extent and its severity. This approach has been validated with real data obtained from myocardial scintigraphy imaging techniques. The obtained results show its effectiveness and validity.

This paper deals with an image segmentation tool inspired from grammar formalism and based on spiral matrix. We are to set a scintigraphic image as a set of lexemes based on a vocabulary of intensities and a set of grammatical rules. Thus, the endocardium, epicardium and epicardial muscle edges are detected. In addition, a set of quantitative information is also deduced such as endocardium area, endocardium diameter and epicardial muscle thickness. This type of task is intended for medical diagnosis assistance. To validate our method, we have segmented a set of real scintigraphic images.

We propose in this work two main interpolation methods (barycentric, spline) applicable to 3D EEG mapping. These 3D methods are used to interpolate scalp potential activity in order to obtain a more efficient spatiotemporal resolution of EEG mapping. Starting from 19 electrodes to generate the potential representations of patients having various behavioral states, we obtain a 3D potential representation of 128 electrodes thanks to the 3D interpolation methods. The evaluation of these algorithms is realized by calculating the normalized Root Mean Squared error (RMS).

Image segmentation is an important task in the image processing and represents a very active research field such as on medical image processing. The aim of this work is to segment an image and make an automated area estimation based on grammar. The entity “language” will be projected to the entity “image” to perform structural analysis and parsing of the image. We achieve to define an image as a set of words based on an alphabet. An object is assimilated as a word recognized by an automaton. We will show how the idea of grammar-based segmentation is applied to synthetic and real problems of cardio-graphic image processing.

This paper presents an optimizing methodology for implementing a multi-layer perceptron (MLP) neural network in a Field Programmable Gate Array (FPGA) device. In order to obtain an efficient implementation, a compromise of time and area is needed. Starting from simulation in the learning phase with fixed point operators, we have developed a methodology which allows the automatic generation of a VHDL code within a multi-width encoding of an MLP. The proposed methodology should determine the optimal encoding of various blocks of our Artificial Neural Networks (ANN) to optimize accuracy and minimize the application area. In addition, real-time constraints should be respected to ensure a reliable classification of vigilance states in humans from electroencephalographic signals (EEG). To validate our approach, the optimized MLP implementation has been tried on Virtex devices.

This paper presents a new technique to generate triangular mesh surface parameterization and characterize 3-D surfaces by invariant spherical harmonic shape descriptors of objects with spherical topology. First, the surface is initially parameterized by defining a continuous one-to-one mapping from the surface of the object to the surface of a unit sphere. Then, the initial parameterization is optimized in a constrained optimization procedure. The obtained parameterized surfaces are expanded into spherical harmonics. The series coefficients are estimated in a least squares sense. Based on harmonic analysis and using results from representation theory, we compute the spherical Fourier transform on the unit sphere S2 with the group of rotations SO(3) as the acting group. The shift theorem allows us to extract invariant 3-D rotation spherical harmonic shape descriptors. The new procedure is illustrated with modelling the left ventricle using the spherical harmonics model and myocardial scintigraphic data. The invariant shape descriptors are used to quantify the heart pathology level.

The main aim in this paper is to study an algorithm of vigilance detection from a minimal number of EEG electrodes, easy to implement on programmable devices, to be used in ambulatory and real everyday life conditions. The connectionist unsupervised approach is summarized in this paper. From the unsupervised classification obtained, a connectionist supervised classification algorithm, the learning vector quantization (LVQ), is used for two different tasks. Firstly, the artefacted states are detected and removed. Secondly, the states deprived of artefacts are then classified in order to decide for the state of vigilance. Connectionist methods with supervised and unsupervised training were used to discriminate the EEG signals characterizing the vigilance states. An artificial neuronal model with a minimal architecture minimizes the complexity and allows implementation. It demonstrates that information, pertinent enough to characterize vigilance states, can be extracted from EEG signal recorded from a single electrode It should also be noted that the intervention of the expert is fundamental in this approach to differentiate nonartefacted vigilance states and artefacted vigilance states.

In this paper, we define the relationship between the location and the degree of the stenosis in coronary arteries and the observed perfusion on the myocardiac scintigraphy. This allows us to model the impact evolution of these stenoses in order to justify a coronarography or to avoid it for patients suspected being in the gray zone. Our approach is decomposed in two steps. The first step consists in modelling a coronary artery bed and stenoses of different location and degree. The second step consists in evaluating the badly perfused volume using only myocardial scintigraphic data.

The authors' aim is to conceive a tool that permits to characterize the levels of vigilance of vehicle drivers by recording in real time the physiological signal variation. The adopted way consists in developing a portable device composed of two parts: the first is analogical for signal shaping and the second is a numerical part. The latter is built up around a DSP (TMS320C31) which ensures signal acquisition from analogue/digital converters, storage, real time treatment and display decision. In this report the authors have implemented an algorithm that permits the treatment and detection of hypovigilance detection. They are currently studying the implementation of a decision algorithm, according to a database of 12 files of 24 hour-EEG registered in volunteers. The authors are studying variations of alpha, beta, theta and delta waves of the EEG in order to define decision levels about transitions of a vigilance state to another.