ICOST Virtual Conference

Alzheimer's disease early detection using a low cost three-dimensional densenet-121 architecture

Abstract: The objective of this work is to detect Alzheimer's disease using Magnetic Resonance Imaging. For this, we use a three-dimensional densenet-121 architecture. With the use of only freely available tools, we obtain good results: a deep neural network showing metrics of 87% accuracy, 87% sensitivity (micro-average), 88% specificity (micro-average), and 92% AUROC (micro-average) for the task of classifying five different classes (disease stages). The use of tools available for free means that this work can be replicated in developing countries.

Self-adaptative Early Warning Scoring System for Smart Hospital

Abstract: With the advent of the Internet of Things (IoT), various interconnected objects can be used to improve the collection and the process of vital signs with partially or fully automatized methods in smart hospital environment. The vital signs data are used to evaluate patient health status using heuristic approaches, such as the early warning scoring (EWS) approach. Several applications have been proposed based on the early warning scores approach to improve the recognition of patients at risk of deterioration. However, there is a lack of efficient tools that enable a personalized monitoring depending on the patient situations. This paper explores the publish-subscribe pattern to provide a self-adaptative early warning score system in smart hospital context. We propose an adaptative configuration of the vital sings monitoring process depending on the patient health status variation and the medical staff decisions.

Machine learning based rank attack detection for Smart Hospital Infrastructure

Abstract: In recent years, many technologies were racing to deliver the best service for human being. Emerging Internet of Things (IoT) technologies made birth to the notion of smart infrastructures such as smart grid, smart factories or smart hospitals. These infrastructures rely on interconnected smart devices collecting real-time data in order to improve existing procedures and systems capabilities. A critical issue in smart infrastructures is the information protection which may be more valuable than physical assets. Therefore, it is extremely important to detect and deter any attacks or breath to the network system for information theft. One of these attacks is the rank attack that is carried out by an intruder node in order to attract legitimate traffic to it. In this paper, we propose an anomaly based rank attack detection system against an IoT network using Support Vector Machines. As a use case, we are interested in the healthcare sector and in particular in smart hospitals which are multifaceted with many challenges such as service resilience, assets interoperability and sensitive information protection. The proposed intrusion detection system (IDS) is implemented and evaluated using Conticki Cooja simulator. Results show a high detection accuracy and low false positive rates.

Remote Health Monitoring Systems based on Bluetooth Low Energy (BLE) Communication Systems

Abstract: Nowadays, remote healthcare monitoring systems (RHMS) are attracting patients, doctors and caregivers. RHMS reduces the number of unessential hospitalizations by providing the required healthcare services for patients at home. Furthermore, continuous health monitoring using RHMS is a hopeful solution for elderly people suffering from chronic diseases. RHMS is in general three tiers architecture where the first tier uses intelligent wearable sensors to gather physiological signs. The majority of wearable sensors constructors commercialized sensing devices with Bluetooth Low Energy (BLE) communication interfaces, which lead to the development of diverse RHMS deploying BLE communication interfaces for physiological patient data gathering. In this paper, we introduce the basic concepts related to RHMS design and development. Besides that, we focus our investigation on the BLE communication protocol used in the healthcare context and its configuration to sense several physiological data. Also, we highlight the different steps enabling reading sensed data on mobile application.

Modeling and specification of Bootstrapping and Registration design patterns for IoT applications

Abstract: The architectures of software systems are becoming more complex, large, and dynamic. The design of these architectures allows architects to master building complex software systems. But, their informal description, may give rise to ambiguity, their understanding becomes more and more difficult and leads to the incorrect implementation of these software systems. There are many solutions allowing software architecture design. In this paper, we use software design patterns as a solution. This is due to their reusable software elements. Our principal objective is to propose other alternatives to the informal visual description of software architectures. In past work, we have studied Service Oriented Architectures. We used SOA design patterns with standard formal notations. This work is a continuation of the past one. We apply our approach on design patterns for the Internet of Things. We introduce a refinement-based approach for modeling IoT design patterns. It takes advantage of graphical modeling and formal method. It is organized around two main axes. The first axis is to provide modeling solutions in conformance with the UML standard language. The second axis covers the specification of design pattern models with the Event-B method and checking the design correctness. Our goal is to design patterns proven correct by construction to successfully apply them and promote their reuse. As a result, we propose a design support tool for IoT architectures based on IoT design patterns. It allows modeling of correct-by-design software systems.

EEG-based Hypo-vigilance detection using convolutional neural network

Abstract: Hypo-vigilance detection provides one of the active research areas in the processing of biomedical signals. For this purpose, electroencephalogram (EEG) is one of the most common modalities in drowsiness and awakeness detection. In this context, we propose a new EEG classication method for detecting fatigue state. Our method makes use of the Convolutional Neural Network (CNN) architecture. We dene an experimental protocol using the Emotiv epoc+ headset. After that, we evaluate our proposed method on a recorded dataset. The reported results show the eciency of our method as compared with other works.

Respiratory Activity Classification Based on Ballistocardiogram Analysis

Abstract: Ballistocardiogram signals describe the mechanical activity of the heart. It can be measured by an intelligent mattress in a totally unobtrusive way during periods of rest in bed or sitting on a chair. The BCG signals are highly vulnerable to artifacts such as noise and movement making useful information like respiratory activities difficult to extract. The purpose of this study is to investigate a classification method to distinguish between seven types of respiratory activities such as normal breathing, cough, and hold breath. We propose a feature selection method based on a spectral analysis namely spectral flatness measure (SFM) and spectral centroid (SC). The classification is carried out using the nearest neighbor classifier. The proposed method is able to discriminate between the seven classes with an accuracy of 94% which shows its usefulness in the context of Telemedicine.

A convolutional neural network for Lentigo diagnosis

Abstract: Using Reflectance Confocal Microscopy (RCM) for lentigo diagnosis is today considered essential. Indeed, RCM allows fast data acquisition with a high spatial resolution of the skin. In this paper, we use a deep convolutional neural network (CNN) to perform RCM image classification in order to detect lentigo. The proposed method relies on an InceptionV3 architecture combined with data augmentation and transfer learning. The method is validated on RCM data and shows very efficient detection performance with more than 98% of accuracy.

Deep Learning-based Approach for Atrial Fibrillation Detection

Abstract: Atrial Fibrillation (AF) is a health-threatening condition, which is a violation of the heart rhythm that can lead to heart-related complications. Remarkable interest has been given to ECG signals analysis for AF detection in an early stage. In this context, we propose an artificial neural network ANN application to classify ECG signals into three classes, the first presents Normal Sinus Rhythm NSR, the second depicts abnormal signal with Atrial Fibrillation (AF) and the third shows noisy ECG signals. Accordingly, we achieve 93.1% accuracy classification results, 92% of sensitivity and 90% of specificity. Furthermore, we yield a value of zero error and a low value of cross entropy, which prove the robustness of the proposed ANN model architecture. Thus, we outperforms the state of the art by achieving high accuracy classification without pre-processing step and without high level of feature extraction, and then we enable clinicians to determine automatically the class of each patient ECG signal.

Unsupervised Method based on Superpixel Segmentation for Corpus Callosum Parcellation in MRI Scans

Abstract: In this paper, we introduce an unsupervised method for the parcellation of the Corpus Callosum (CC) from MRI images. Since there are no visible landmarks within the structure that explicit its parcels, non-geometric CC parcellation is a challenging task especially that almost of proposed methods are geometric or data-based. In fact, in order to subdivide the CC from brain sagittal MRI scans, we adopt the probabilistic neural network as a clustering technique. Then, we use a cluster validity measure based on the maximum entropy (Vmep) to obtain the optimal number of classes. After that, we obtain the isolated CC that we parcel automatically using SLIC (Simple Linear Iterative Clustering) as superpixel segmentation technique. The obtained results on two challenging public datasets prove the performance of the proposed method against geometric methods from the state of the art. Indeed, as best as we know, it is the first work that investigates the validation of a CC parcellation method on ground-truth datasets using many objective metrics.

Using learning techniques to observe elderly's behavior changes over time

Abstract: Smart environments and technology used for elder care, increases independent living time and cuts long-term care costs. A key requirement for these systems consists in detecting and informing about abnormal behavior in users'routines. In this paper, our objective is to automatically observe the elderly behavior over time and detect anomalies that may occur on the long term. Therefore, we propose a learning method to formalize a normal behavior pattern for each elderly people related to his Activities of Daily Living (ADL).We also adopt a temporal similarity score between activities that allows to detect behavior changes over time. In change behavior period we focus on each activity to detect anomalies. A use case with real datasets are promising.

Personalized and Contextualized Persuasion System for Older Adults’ Physical Activity Promoting

Abstract: Aging often involves a significant change in roles and social positions. The greatest health risk for seniors is the adoption of a sedentary lifestyle that causes isolation, depression and many diseases. However, convincing an older adult to regularly d physical activities isn’t generally a simple mission. This paper proposes a personalized and contextualized persuasion system to promote physical activities for older adults. In fact, our approach considers the personal and health profile of the older adult. It also takes into account different context parameters (context-awareness). This intelligence is guaranteed thanks to the use of the semantic modeling and reasoning, which from different types of information would be able to decide the best moment to trigger notifications from our persuasive system to the participating older adults.

Baseline Modelling and Composite Representation of Unobtrusively (IoT) Sensed Behaviour Changes Related to Urban Physical Well-being

Abstract: We present the grounding approach, deployment and preliminary validation of the elementary devised model of physical well-being in urban environments, summarizing the heterogeneous personal Big Data (on physical activity/exercise, walking, cardio-respiratory fitness, quality of sleep and related lifestyle and health habits and status, continuously collected for over a year mainly through wearable IoT devices and survey instruments in 7 global testbed cities) into 5 composite domain indicators/indexes convenient for interpretation and use in predictive public health and preventive interventions. The approach is based on systematized comprehensive domain knowledge implemented through range/threshold-based rules from institutional and study recommendations, combined with statistical methods, and will serve as a representative and performance benchmark for evolution and evaluation of more complex and advanced well-being models for the aimed predictive analytics (incorporating machine learning methods) in subsequent development underway.

Automatic Daily Activity Schedule Planning for Simulating Smart House with Elderly People Living Alone

Abstract: A simulation tool that supports developers to build scenarios automatically in multiple simulation platforms is proposed. As an essential part of this simulator, this study proposed an activity schedule generator to mimic the daily life of elderly people living alone. This generator outperforms existing methods of activity schedule planning in three aspects: 1) it is adaptive to the layout of a simulated smart house; 2) there is no unspecified time in the timeline of generated schedules; and 3) it generates stable, but not tedious schedules for a number of days. A real-time location data generator is proposed to convert generated schedules to simulated real-time location data of the resident, and a proposed interface converts these simulated location data to simulated records of virtual passive infrared (PIR) sensors, which can be used to optimize placement of PIR sensors in a smart house.

A novel on-wrist fall detection system using Supervised Dictionary Learning technique

Abstract: Wrist-based fall detection system provides a very comfortable and multi-modal healthcare solution especially for elderly risking falls. However, the wrist location presents a very challenging and unstable spot to distinguish falls among other daily activities. In this paper, we propose a Supervised Dictionary Learning approach for a wrist-based fall detection. Three Dictionary learning algorithms for classification are invoked in this study, namely SRC, FDDL, and LRSDL. To extract the best descriptive representation of the signal data we followed different preprocessing scenarios based on accelerometer, gyroscope, and magnetometer. A considerable overall performance was obtained by the SRC algorithms reaching respectively 99.8\%, 100\% and 96.6\% of accuracy, sensitivity, and specificity using raw data provided by a triaxial accelerometer, accordingly overthrowing previous proposed methods for wrist placement.

Improving Access and Mental Health for Youth through Virtual Models of Care

Abstract: The overall objective of this research is to evaluate the use of a mobile health smartphone application (app) to improve the mental health of youth between the ages of 14-25 years, with symptoms of anxiety/depression. This project includes 115 youth who are accessing outpatient mental health services at one of three hospitals and two community agencies. The youth and care providers are using eHealth technology to enhance care. The technology uses mobile questionnaires to help promote self-assessment and track changes to support the plan of care. The technology also allows secure virtual treatment visits that youth can participate in through mobile devices. This longitudinal study uses participatory action research with mixed methods. The majority of participants identified themselves as Caucasian (66.9%). Expectedly, the demographics revealed that Anxiety Disorders and Mood Disorders were highly prevalent within the sample (71.9% and 67.5% respectively). Findings from the qualitative summary established that both staff and youth found the software and platform beneficial.

Combined Machine Learning and Semantic Modelling for Situation Awareness and Healthcare Decision Support

Abstract: The average of global life expectancy at birth was 72 years in 2016, however, the global healthy life expectancy at birth was only 63.3 years in the same year, 2016. Living a long life is not any more as challenging as assuring active and associated life. We propose in this paper an IoT based holistic remote health monitoring system for chronically ill and elderly patients. It supports smart clinical decision help and prediction. The patient heterogeneous vital signs and contexts gathered from wore and surrounding sensors are semantically simplified and modeled via a validated ontology composed by FOAF, SSN/SOSA and ICNP ontologies. The reasoner engine is based on a scalable set of inference rules cohesively integrated with an ML algorithm to ensure predictive analytic and preventive personalized health services. Experimental results prove the efficiency of the proposed system.

Study of Middleware for Internet of Healthcare Things and their applications

Abstract: The rapid proliferation and miniaturization of the wireless and embedded devices has led to the invasion of the Internet of Things in many domains and has reached the healthcare sector to form what is called the Internet of Healthcare Things (IoHT). The growing number of the applications in the Internet of Healthcare Things as well as the overwhelming number of heterogeneous medical devices that should interact in this network has put the researches and developers in front of lot of challenges : How to facilitate the implementation of the various healthcare applications ? And how to ease the integration of new devices and make their interoperation a transparent task for the developers ? To fulfill these requirements, lot of middleware have been proposed. In this paper we provide a complete study on the existing middleware for IoHT and specify their applications, we propose a taxonomy for them and we present their main advantages and drawbacks.

Uncertainty in IoT for smart Healthcare : Challenges, and Opportunities

Abstract: According to F.H. Knight, uncertainty signifies deviations from the expected states, which prevent us from the use of any probability for the determination of a result for a given action or decision [1]. This paper describes the phenomenon of uncertainty in the face of technological megatrends and challenges associated with them. The article focuses on the analysis of the uncertainty in one of the most important technology trends – the Internet of Things (IoT) – on the example of Healthcare. The right decisions are not always equivalent to good results. Sometimes, the decision taken in accordance with general rules brings worse results than the one who breaks them. Such a situation is possible as a result of the uncertainty accompanying the predictions of the future. In this article the concept of the IoT is treated as a big, complex, dynamic system with specific characteristics, dimensions. structures and behaviors. The aim of the article is to analyze the factors that may determine the uncertainty and ambiguity of such systems in the context of the development of Healthcare, and recommendations are made for future research directions

Secure E-health Platform

Abstract: Currently, the Internet has become a service hosting infrastructure through its interconnection of a very large number of heterogeneous objects, thus offering users several types of services implemented by different sectors. Although these services make people’s lives easier and provide them with a means of communication between their real and virtual worlds, they risk being a path of intrusion into their private lives, or in some cases an easy target for malicious individuals aiming to endanger human life. To avoid this, we have designed a secure e-health platform based on IoT that serves to monitor patients’ medical profiles remotely by collecting their medical records while ensuring their confidentiality and integrity.

Hybrid and Secure E-health Data Sharing Architecture in Multi-Clouds Environment

Abstract: Healthcare is among the sectors showing efforts in adopting cloud computing to its services considering the provided cost reduction and healthcare process efficiency. However, outsourcing patient's sensitive data increases the concerns regarding security, privacy, and integrity of healthcare data. Therefore, there is a need for building a trust relationship between patients and e-health systems. Hence, several security measures should be taken to achieve the mentioned goal. In this paper, we propose a privacy-preserving framework, called Hybrid and Secure Data Sharing Architecture (HSDSA), to secure data storage in e-health systems. Our approach improves security in healthcare by maintaining the privacy and confidentiality of sensitive data and preventing threats. In fact, in the upload phase, Multi-cloud environment is used to store Rivest–Shamir–Adleman (RSA) encrypted medical records. We adopt a Shamir's secret sharing approach for the distribution of shares to different independent cloud providers. In the retrieval phase, the reconstruction operation is based on the (t, n) strategy. To check the requester identity and to prove the hash possession, we used a zero-knowledge cryptography algorithm, namely the Schnorr algorithm. The patient has a total control over the generation and management of the decryption keys using Diffie-Hellman algorithm without relying on a trusted authority.

Blockchain for Internet of Medical Things: A technical review

Abstract: The Internet of Medical Things (IoMT) represents a network of implantable or wearable medical devices that continuously collect medical data about the patient's health status. These data are heavy, sensitive and require high level of security. With the emergence of blockchain technology, researchers are focusing on using blockchain strategies to bring security to healthcare applications. However, such integration is very difficult and challenging due to the different requirements in these two technologies. We present in this paper a technical review of existing solutions applying blockchain technology on IoMT. We analyze these studies, discuss the proposed architectures and how they managed the integration challenges. The open issues regarding the application of blockchain over IoMT are also specified.

Application of Blockchain Technology inHealthcare: A comprehensive study

Abstract: Blockchain technology has been emerged in the last decade and has gained a lot of interests from several sectors such as finance, government, energy, health, etc. This paper gives a broad ranging survey of the application of blockchain in healthcare domain. In fact, the ongoing research in this area is evolving rapidly. Therefore, we have identified several use cases in the state of art applying the blockchain technology, for instance for sharing electronic medical records, for remote patient monitoring, for drug supply chain, etc. We have focused also on identifying limitations of each studied approach and finally we have discussed some open research issues and the areas of future research.

Trust Execution Environment and Multi-party computation for Blockchain e-health systems

Abstract: Blockchain is a rich and attractive domain for researchers since it is independent of "third party" such as Bank or government. This "open" phenomenon does not respect all the security criteria such as private data protection and confidentiality; hence, we cannot trust this approach despite its contributions. Blockchain technology has gained considerable progress in recent years in fields such as e-health. The medical data contains personal and sensitive information that must be preserved. The current Blockchain systems suffer from serious practical limitations, e.g. poor performance, high-energy consumption and lack of confidentiality. On the other hand, Trust Execution Environment TEE is imperfect; it is based on the centralization of data. To avoid data centralization and its limitations, an approach based on collecting the necessary data from distributed database is presented in this paper. Our goals are to protect the user's privacy and to execute it in TEE combined with Multi-party computation MPC. We proof by security analysis that our new solution meets the fundamental criteria of security such as confidentiality and privacy.

A fuzzy-ontology based diabetes monitoring system using internet of things

Abstract: The number of people suffering from chronic diseases is rapidly increasing, providing the healthcare services industry more challenging problems. Recently, several health monitoring systems based on Internet-of-things (IoT) have been adopted to continuously supervise the chronic patients ‘health and improve the quality of healthcare services. The majority of these systems are knowledge-based and use ontologies to collect, integrate, and interoperate IoT data. It is quite noted that classical ontologies cannot appropriately treat imprecise and ambiguous knowledge for healthcare applications, but fuzzy ontology can effectively resolve data and knowledge problems with uncertainty. This paper presents a novel fuzzy-ontology based system that integrates the internet of things technologies. The proposed system allows continues health monitoring of diabetic patients. It collects data from wearable and environmental sensors, extracts the values of patient risk factors, determines the patient's health condition and generates personalized recommendations. The combination of Fuzzy Logic and ontology considerably raises the prediction accuracy of a patient's health condition and the average precision of healthcare decisions and recommendations. In this work, a fuzzy ontology-based model is developed using the Protégé tool. Additionally, a semantic fuzzy decision-making mechanism using the proposed fuzzy-ontology is also developed based on the Semantic Web Rule Language (SWRL) rules and fuzzy logic. The ontology is evaluated using the Simple Protocol and RDF Query Language (SPARQL) queries based on realistic scenarios of patient monitoring. The results indicate the feasibility of the system for effective remote continuous monitoring.

A Hybrid Approach for Heart Disease Diagnosis and Prediction Using Machine Learning Techniques

Abstract: Heart disease is considered as one of the major causes of death throughout the world. It cannot be easily predicted by the medical practitioners as it is a difficult task which demands expertise and higher knowledge for prediction. Currently, the recent development in medical supportive technologies based on data mining, machine learning plays an important role in predicting cardiovascular diseases. In this paper, we propose new hybrid approach to predict cardiovascular disease using different machine learning techniques such as Logistic Regression (LR), Adaptive Boosting (AdaBoostM1), Multi-Objective Evolutionary Fuzzy Classifiers (MOEFCs), Fuzzy Unordered Rule Induction (FURIA), Genetic Fuzzy System-LogitBoost (GFS-LB) and Fuzzy Hybrid Genetic Based Machine Learning (FH-GBML). For this purpose, the accuracy and results of each classifier have been compared, with the best classifier chosen for a more accurate cardiovascular prediction. With this objective, we use two free software (Weka and Keel).

Context-aware healthcare adaptation model for COPD diseases

Abstract: Nowadays, ubiquitous computing and mobile applications are controlling all our life’s aspects, from social media and entertainment to the very basic needs like commerce, learning, government, and health. These systems have the ability to self-adapt to meet changes in their execution environment and the user’s context. In the healthcare domain, information systems have proven their efficiency, not only by organizing and managing patients’ data and information but also by helping doctors and medical experts in diagnosing disease and taking precluding procedure to avoid serious conditions. In chronic diseases, telemonitoring systems provide a way to monitor the patient’s state and biomarkers within their usual life’s routine. In this article, we are combining the healthcare telemonitoring systems with the context awareness and self-adaptation paradigm to provide a self-adaptive framework architecture for COPD patients.

Study of healthcare professionals’ interaction in the patient records with annotations

Abstract: HThe annotation practice is an almost daily activity; it is used by healthcare professionals (PHC) to analyze, collaborate, share knowledge and communicate, between them, information present in the healthcare record of patients. These annotations are created in a healthcare cycle that consists of: diagnosis, treatment, advice, follow-up and observation. Due to an exponential increase in the number of medical annotation systems that are used by different categories of health professionals, we are faced with a problem of lack of organization of medical annotation systems developed on the basis of formal criteria. As a result, we have a fragmented image of these annotations tools which make the mission of choice of an annotation system by a PHC, in a well-defined context (biology, radiology…) and according to their needs to the functionalities offered by these tools, are difficult. In this article we present a classification of thirty annotation tools developed by industry and academia based on 5 generic criteria. We conclude this survey paper with model proposition.

Multirate ECG Processing and k-Nearest Neighbor Classifier Based Efficient Arrhythmia Diagnosis

Abstract: The goal of this work is to make a contribution to the development of a wearable smart and computationally efficient multirate Electrocardiogram (ECG) automated detector of arrhythmias. It utilizes an intelligent combination of multirate denoising plus wavelet decomposition for an effective realization of the ECG wearable processing chain. The decomposed signal subband features are mined and in next step these are utilized by the mature k-Nearest Neighbor (KNN) based classifier for an effective arrhythmia diagnosis. The multirate feature diminishes the system processing load. It induces a substantial reduction in the processing activity of the system and thus allows a substantial decrease in energy consumption compared to traditional counterparts. The overall performance of the system is estimated in terms of the accuracy of the classification process. Obtained results reveal an overall 22.5 times compression gain and 4 folds processing outperformance of the devised approach over the traditional equals while securing 93.2% highest classification accuracy. Findings confirm that the proposed solution could potentially be embedded in contemporary automatic and mobile cardiac diseases diagnosis systems.

Comparative Study of Relevant Methods for MRI/X Brain Image Registration

Abstract: Brain image registration is a challenging problem in medi-cal imaging that needs to be studied. Several methods of brain imageregistration have been proposed in order to overcome the requirement ofclinicians. In this paper, we assess the performance of a hybrid method forbrain image registration against standard registration tools. Most tradi-tional registration tools use different methods for mono- and multi-modalregistration, whereas the hybrid registration method is providing bothmono and multi-modal brain registration of PET, MRI and CT images.To determine the appropriate registration method, we used two brain im-age datasets (CAHM and RIRE), as well as two evaluation metrics (NMIand NCCC). From the extensive set of experiments that we performed,we can deduce that the hybrid method based on adaptive mutual infor-mation coupled with curvelet coefficients outperforms all other standardregistration tools and has achieved promising registration accuracy.

Machine learning Classification models with SPD/ED Dataset: Comparative study of Abstract versus full article approach

Abstract: In response to the researchers need in the bio-medical domain, we opted for automating the bibliographic research stage. In this context, several classification models of supervised machine learning are used. Namely the SVM, Random Forest, Decision Tree, KNN, and Gradient Boosting. In this paper, we conduct a comparative study between experimental results of full article classification and abstract classification approaches. Furthermore, we evaluate our results by using evaluation metrics such as accuracy, precision, recall and F1-score. We observe that the abstract approach outperforms the full article approach in terms of learning time and efficiency. Finally, we will outline our future work.

Evaluation of stationary wavelet transforms in reconstruction of pure high frequency oscillations (HFOs)

Abstract: High frequency oscillations (HFO) from, MEG (magnetoencephalography) and intracerebral EEG are considered as effective tools to identify cognitive status and several cortical disorders especially in the diagnosis of epilepsy. The aim of our study is to evaluate stationary wavelet transform (SWT) technique performance in efficient reconstruction of pure high frequency epileptic oscillations, reputed biomarkers of epileptogenic zones: generators of interictal epileptic discharges, and offhand seizures. We applied SWT on simulated and real database to detect non-contaminated HFO by spiky element. For simulated data, we computed the GOF of reconstruction that reaches for all studied constraint (relative amplitude, frequency, SNR and overlap) very promising results. For real data we used time frequency domain to evaluate the robustness of SWT reconstruction of HFO. We proved that SWT is an efficient filtering technique for separation HFO from spiky events. Our results would have an important impact on the definition of epileptogenic zones.

Ensuring the Correctness and Well Modeling of Intelligent Healthcare Management Systems

Abstract: Recent research focus more and more on IoT systems and their applications in order to make people life easier and controllable. The main aim is to expand IoT applications and services into various domains while ensuring communication and automated exchange between them. Recent research handles many issues related to IoT especially implementation, modeling, and deployment. However, many challenges need more deep and thorough analysis especially in terms of flexible modeling, extensible implementation, with respect to the privacy issue. This work focuses principally on modeling IoT systems dedicated to smart healthcare case. We attempt to address the emergency service by initiating a modeling mechanism for Healthcare Management System (HMS) by using UML diagrams, and propose an appropriate access control in order to reinforce it. Then, we ensure the correctness of the developed HMS by relying on the verification and validation based on a formal analysis that showed significant results by using Alloy tool.

An Embedded ANN Raspberry PI for Inertial Sensor Based Human Activity Recognition

Abstract: Human Activity Recognition (HAR) is one of the critical subjects of research in health and human machine interaction fields in recent years. Algorithms such as Support Vector Machine (SVM), K-Nearest Neighbors (K-NN), Decision Tree (DT) and many other algorithms were previously implemented to serve this common goal but most of the traditional Machine learning proposed solutions were not satisfying in term of accuracy and real time testing process. For that, a human activities analysis and recognition system with an embedded trained ANN model on Raspberry PI for an online testing process is proposed in this work. This paper includes a comparative study between the Artificial Neural Network (ANN) and the Recurrent Neural Network (RNN), using signals produced by the accelerometer and gyroscope, embedded within the BlueNRG-Tile sensor. After evaluate algorithms performance in terms of accuracy and precision which reached an accuracy of 82% for ANN and 99% for RNN, obtained ANN model was implemented in a Raspberry PI for real-time predictions. Results show that the system provides a real-time human activity recognition with an accuracy of 86%.

Human Activities Recognition in Android Smartphone using WSVM-HMM Classifier

Abstract: Being able to recognize human activities is essential for several applications such as health monitoring, fall detection, context-aware mobile applications. In this work, we perform the recognition of the human activity based on the combined Weighted SVM and HMM by taking advantage of the relative strengths of these two classification paradigms. One significant advantage in WSVMs is that, they deal the problem of imbalanced data but his drawback is that, they are inherently static classifiers - they do not implicitly model temporal evolution of data. HMMs have the advantage of being able to handle dynamic data with certain assumptions about stationary and independence. The experiment results on real datasets show that the proposed method possess the better robustness and distinction.

Mobile Assistive Application for Blind People in Indoor Navigation

Abstract: Navigation is an important human task that needs the human sense of vision. In this context, recent technologies developments provide technical assistance to support the visually impaired in their daily tasks and improve their quality of life. In this paper, we present a mobile assistive application called "GuiderMoi" that retrieves information about directions using color targets and identifies the next orientation for the visually impaired. In order to avoid the failure in detection and the inaccurate tracking caused by the mobile camera, the proposed method based on the CamShift algorithm aims to introduce better location and identification of color targets. Tests were conduct in natural indoor scene. The results depending on the distance and the angle of view, defined the accurate values to have a highest rate of target recognition. This work has perspectives for this such as implicating the augmented reality and the intelligent navigation based on machine learning and real-time processing.

Older People’s Needs and Opportunities for Assistive Technologies

Abstract: Older adults experience a disconnect between their needs and adoption of technologies that have potential to support more independent living. This pa-per reviewed research that links people’s needs with opportunities for assis-tive technologies. It searched 13 databases identifying 923 papers with 34 papers finally included for detailed analysis. The research papers identified needs in the fields of health, leisure, living, safety, communication, family re-lationship and social involvement. Amongst these, support for activities of daily living category was of most interest. In specific sub-categories, the next most reported need was assistive technology to support walking and mobility followed by smart cooking/kitchen technology and assistive technology for social contacts with family member/other people. The research aimed to in-form a program of research into improving the adoption of technologies where they can ameliorate identified needs of older people.

Towards a Formal Context-Aware Workflow Model for Ambient Environment

Abstract: Ambient systems owns some particular characteristics that makes their context awareness a sincere problem; they are composed of heterogeneous distributed devices, some of these devices may appear and disappear during operations. In addition, users interacting in these systems are themselves dynamic. Therefore, context-aware workflow management allows workflows to adapt dynamically according to the environment changes. Context information are complex and diverse which makes the modeling the key issue. This paper presents an approach to model context-aware workflows. First, we describe the workflow using Ag-LOTOS. Then, based on this description, we build the contextual planning system CPSw that allows the presentation of the context at each activity state.

The PULSE project: a case of use of big data uses toward a cohomprensive health vision of city well being

Abstract: Despite the silent effects sometimes hidden to the major audience, air pollution is becoming one of the most impactful threat to global health. Cities are the places where deaths due to air pollution are concentrated most. In order to correctly address intervention and prevention thus is essential to assest the risk and the impacts of air pollution spatially and temporally inside the urban spaces. PULSE aims to design and build a large-scale data management system enabling real time analytics of health, behaviour and environmental data on air quality. The objective is to reduce the environmental and behavioral risk of chronic disease incidence to allow timely and evidence-driven management of epidemiological episodes linked in particular to two pathologies; asthma and type 2 diabetes in adult populations. developing a policy-making across the domains of health, environment, transport, planning in the PULSE test bed cities

ForeSight - an AI-driven Smart Living Platform, Approach to Add Access Control to openHAB

Abstract: We have created an approach for a smart living platform called ForeSight which consists of different modules: a service engineering module, a Web of Things (WoT)-based Internet of Things (IoT) module and an artificial intelligence (AI) component. This paper describes how openHAB, a smart home middleware is extended to fulfil platform requirements related to a successful interaction with the IoT module of ForeSight, more precisely: to add identity and access management (IAM) to openHAB. European privacy laws need to be considered.

PULSE Project Overview

PULSE is a H2020 project aimed to improve predictive analytics and risk detection in cities as well as influence lifestyles and behaviour. The project analyses the environmental and behavioural determinants of disease onset by focusing in particular on the link between air pollution and asthma, and between physical inactivity and Type 2 Diabetes. Moreover, the project is addressing community wellbeing and its relation with health outcomes. The final goal is to build extensible models and technologies to predict, mitigate and manage public health problems and implement a Public Health Observatory which will serve as linked hub that utilises knowledge-driven processes and big data to shape intersectorial public policies and service provision.

PULSE Technical Architecture

PULSE proposes three technological components :

• An app for citizens, which aims to collect both subjective and objective data to assess the user's habits, psychosocial status and health risk (asthma and Type 2 diabetes). The app provides supportive feedback, educational materials, and air quality awareness updates.
• A dashboard for public health organizations designed to show aggregated population statistics and assess health risks. The system also provides advanced tools to perform simulations of population profiles, air quality levels and advanced geo-spatial analytics.
• A Community of Practice, a website to share the PULSE experience of the seven PULSE pilots cities.

• Jo Ivey Boufford, Clinical Professor of Global Health (NYU School of Global Public Health)
• Thomas Kirchner, Director and Principal Investigator of the mHealth (mobile health) Lab (NYU School of Global Public Health)
• Elisabeth Bengsston, Consultant (WHO Healthy Cities Network, Division of Policy and Governance for Health and Well-being)
• Fiona Donovan, Project Manager (National Healthy Cities and Counties of Ireland at Health Service Executive)

PULSE (Participatory Urban Living for Sustainable Environments) aims to transform public health by leveraging the power of Big Data technology, mobile, sensor devices and Artificial Intelligence. The project collaborates with seven smart cities (Paris, Barcelona, Birmingham, New York City, Singapore, Pavia and Taipei) that provide open city data, data from clinical records, environmental sensors, personal devices and health trackers to assess the health and wellbeing at individual and community level. Specifically, the project focuses on Type 2 Diabetes and asthma in the adult population.The project uses data on modifiable and non-modifiable biological, behavioral, social and environmental risk factors to create personalized risk stratification models. The project fosters the development of a knowledge-based and data-driven approach to refine the concept of urban well-being and align it with recommendations from the World Health Organization (WHO). PULSE also aims to promote the adoption of Health in all Policies (HiaP) by integrating information from neighborhoods, environments, transportation and urban planning in each city. To do so, PULSE is developing simulation tools to assist decision makers in urban planning and health promotion. To better understand the project, we invite you to watch an introductory video.

In this demo session, we will assess three technological components of the PULSE Project:

• PULSAIR: An app for citizens, which aims to collect both subjective and objective data to assess the user's habits, psychosocial status and health risk (asthma and Type 2 diabetes). The app provides supportive feedback, educational materials, and air quality awareness updates.
• A PHO Dashboard: A dashboard for public health organizations designed to show aggregated population statistics and assess health risks. The system also provides advanced tools to perform simulations of population profiles, air quality levels and advanced geo-spatial analytics.
• A Community of Practice: a website to share the PULSE experience of the seven PULSE pilots cities.

Pulsair is the PULSE app intended for use by everyday community members. It has been developed and published in the Apple App and Google Play stores. The Pulsair app aims to foster healthy lifestyles and to increase awareness of air pollution hotspots. The app can be connected to FitBit, Garmin and Asus health tracker devices and community members can provide subjective data (from questionnaires embedded in the app), physiological and activity data (from the wearable sensors) and GPS data. The app provides information on health risks and behavioral modifications that can mitigate those risks. Furthermore, the app combines location data from GPS and environmental sensors (e.g. PM 2.5, CO, etc) to help users understand their exposure to air pollution. By using Pulsair, the user also shares data back with local public health agencies, which can then be used to assess neighbourhoods and understand the health needs of the population.

To register and configure the activity tracker, first, the user must install and configure the app. Each user receives a unique invitation code that can be used to access the registration process. Next, the app prompts users to complete their profile with demographic information. At the end of the registration process, users will connect their activity tracker (Fitbit, Asus, or Garmin) to the app.

The app periodically presents users with questions to gather information about their health and lifestyle habits. Once the system has collected all the required information, it assesses the user’s risk of developing Type 2 diabetes in 10 years. After this risk score is generated, the app provides the user with feedback and recommended educational content to help users modify their health behaviors and ultimately reduce their risk of developing Type 2 diabetes.

The app periodically presents users with questions to gather information about their health and lifestyle habits. Once the system has collected all the required information, it assesses the user’s risk of developing Type 2 diabetes in 10 years. After this risk score is generated, the app provides the user with feedback and recommended educational content to help users modify their health behaviors and ultimately reduce their risk of developing Type 2 diabetes.

The Pulsair app also gathers psychosocial information from users to assess wellbeing at the individual level. The app does this by using a model with 6 dimensions:

• Life satisfaction
• Emotional and mental health
• Daily functioning and ability to complete with daily tasks
• Physical energy level
• Strength of community and support system
• Strength of interpersonal relationships involving trust and helping one another

The app generates a score for each of the dimensions and provides personalized feedback and modifications that can be made to increase overall wellbeing. In the example scenario presented in the video, the app offers suggestions for users that reported having a bad mood and poor sleep quality.

The Pulsair app’s "city" functionality allows users to check current air quality measurements in their city. The app can connect to the PULSE air quality sensor network or to specific internet services that provide air quality measurement estimates based on the user’s location and the city in which the user lives.

In addition, users can also receive an estimate of their particle ingestion (e.g., PM2.5) during an outdoor activity. To do this, users track their outdoor activity by submitting the start time and end time of their activity. After an hour, the user can return to the app and will be able to see a map with exposure results and a summary of the estimated particle ingestion.

This is the landing page for the dashboards where the system presents the overall data collected from the city. This home page presents the top statistics on Type 2 diabetes risk, asthma risk, and wellbeing levels. In the right corner of the page, viewers can see overall air quality levels of the cities that have been included in the Pilot portion of this project. A map of each city shows the number of Pulsair users as well as activity trackers statistics (e.g., average # of steps collected by a FitBit).

The user can select a specific risk assessment model (Type 2 diabetes or asthma) and see detailed statistics on the health condition’s risk models.

The urban health analytics functionality shows all the data that have been gathered through the Pulsair app. On this page, users can explore sociodemographic, lifestyle, health status (mental and physical) and the neighborhood characteristics reported by app users.

Using the home page, viewers can see the historic and statistical air pollution data. The user can select the available stations (from the PULSE sensor network and from imported data) and see the evolution of air pollution.

The PULSE Community of Practice (CoP) aims to generate a collaborative web platform that includes knowledge and experiences acquired from the pilot users in the seven testbed cities. With this information, the PULSE CoP hopes to create opportunities for community members, municipalities and other stakeholders to upload data, provide feedback, disseminate information, exchange best practices and examples of research, and share use cases within the project.

The public page shows the experience of the PULSE test beds, a description of the pilot, an overview of the participants and other city initiatives, and recommended media channels for following the city’s activities related to health promotion.