International audience

The advent of real-time data processing applications has revolutionized monitoring and analysis capabilities across diverse operational domains, including big data and healthcare. This study focuses on the development and evaluation of a prototype application designed for real-time data processing in dynamic environments. The application parses configuration files, processes real-time sensor data and outputs calculated values at user-defined intervals, ensuring accuracy and timeliness in data analysis. Experimental evaluations under varied conditions validate the application’s robust performance in managing pipeline operations, computational efficiency, response times and data aggregation precision. Optimizations, including data type adjustments, significantly enhance network communication efficiency and reduce latency, critical for supporting real-time applications. The application’s flexibility in user-configured settings for data storage and aggregation proves essential for adapting to specific application requirements. This adaptability is particularly beneficial f or handling the complexities of big data and the sensitivity of healthcare data. Overall, this study contributes a robust solution for real-time data processing and continuous monitoring, demonstrating its feasibility and applicability across dynamic operational environments, including those requiring rigorous data handling and precise analytics.

Due to the health data privacy issues, wearable devices are less useful in the industry and can not reflect their potential power. Besides, wearable health devices bring constraints such as limited energy budget, needs to recharge and impact on daily activities. To recover these problems and fulfill healthcare needs, the recent trends in the literature is to use wireless signal to capture/recognize human activities. When human activity and emotion recognition study fields merging with the machine learning techniques, can offer a innovative systems. In recent years, with the help of wireless sensing mechanisms, this study field have gained more importance. So that, these wireless sensing mechanisms have eliminated the requirement of physical body contact such as sensor devices. Nowadays, wireless sensing technologies are preferred with the easy and cost-effective access to off-the-shelf WiFi-enabled devices. In this research, a human activity and emotion recognition system have proposed with usage of WiFi channel state information(CSI) mechanism that modelled by machine learning algorithms. As a result, with a tolerable error rate, it is possible to distinguish person’s heart rate information with a contactless solution via wi-fi signal.

Respiratory rate (RR) is a physiological parameter typically used to monitor patient status in clinical settings. The goal of the Respiratory Rate Prediction Project is to use supervised machine learning techniques to estimate a person’s respiratory rate using real-time, continuous Photoplethysmogram (PPG) and Electrocardiogram (ECG) and oximeter data. In addition, it is also our goal to investigate the feasibility of using such data to improve diagnostic processes in healthcare. It consists of a series of studies of different algorithms for respiratory rate estimation from clinical data and is complemented by the provision of publicly available datasets and resources.

With the image processing techniques, nowadays it is possible to obtain 3D X-ray tomography which is used in medical diagnosis. Such systems require a reconstruction of an object - a human body in this case- in 3D from a set of its 2D projections. The reconstruction volume is usually discretized on a regular grid of isotropic voxels which implies an increase in their number to achieve good spatial resolution. In this work, we propose a process to obtain 3D mesh generation. method allowing to discretize the 3D reconstruction space in a relevant way directly from the structural information contained in the projection data. The idea is to obtain a representation adapted to the studied object. Here we have recourse to a tetrahedral mesh matching the structure of the object: the mesh density adapts according to the interfaces and homogeneous regions. To build such a mesh, the first step of the method consists in detecting the edges in the 2D projection data. The structural information thus obtained is then merged in the reconstruction space in order to construct a point cloud sampling the 3D interfaces of the imaged object.

The correct and flawless functioning of medical devices is of utmost importance for patients and the medical sector. Although medical device manufacturers utilize various protocols and procedures during the development and production of medical devices, the integrated circuits are usually manufactured by other parties. During the manufacture of the semiconductors utilized in the integrated circuits, the performance of the components can vary from batch to batch and even samples within the same batch can have different performance characteristics. For applications in life critical applications the selection of the most fit for purpose components is vital. In this paper, we propose a novel method for the semiconductor industry to verify the quality of transistors. The I-V graphs of the components are evaluated by multiple Convolutional Neural Networks (CNN) using visual data similar to expert evaluation, then these Machine Learning (ML) architectures utilize a multi model ensemble technique where one architecture providing a negative output will overrule the vote of the other architectures is utilized to assure very stringent quality control. Our method is tested on CMOS transistors and the results are comparable to those of experts with 10 years of experience in the industry.

Wearing a non-medical mask or face covering helps reduce the spread of COVID-19 in the community. The use of non-medical masks or face covers changes communication for everyone, but this presents an additional challenge for people with hearing loss or communication difficulties. People with hearing loss may have difficulty hearing in difficult situations, such as in noisy places or when they cannot trust lip reading cues or body expressions. face. In this study, a survey is implemented where speech recognition level-based questions are asked to the participants. Then, the same questions are asked again for the cases where medical masks are worn. In doing so, participants are also asked to explain the main difference between the pandemic period and the time before pandemic. Results show that wearing a medical mask has a concrete impact on speech recognition and causes a degrade on hearing level.

Air pollution is the contamination of the internal or external environment of any chemical, physical or biological material and altering the natural properties of the atmosphere. Air pollution is caused by natural causes such as volcanic eruptions, forest fires, or excessive evaporation, as well as by increasing traffic density, especially by artificial causes such as fuel types used by urban inhabitants and industrialization. Each year, countries keep statistics of deaths caused by different types of direct air pollution or indirectly by diseases caused by these different types of pollution. In this study, the impact of countries' geographical region and population density on mortality rates due to air pollution is analyzed. A dataset of Turkey and the twenty six countries covers information such as the population density, different types of pollution rates, number of deaths caused by each type of pollution and their ratios to current population number throughout a time period between 1990-2017. In the data analysis, air pollution types were examined one by one and it was sought to identify the similarities among countries. As a result of these analysis, it was shown that there are similarities between Turkey and the different neighbour countries on the death cases observed due to different types of air pollution.

Investigating the incidence, type, and preventability of adverse drug events (ADE) and medication errors is crucial to improving the quality of health care service. ADEs, medications errors can be extracted from practice data, patients feedback and especially from medication order. In this study, we examine the dataset filled with medication orders and with a web service based approach, each medication order is analyzed to avoid from drug-drug interaction. Firstly, the study is performed on a training set to explore the drug-drug interactions based on the clinical drug component of each medication. Then the proposed web service approach is integrated into the actual centralized Medication Order Management System (MOMS) of the hospital. With this real time checking ADE mechanism, doctors are warned in case there is a possible ADE or a medication error between the medications.