Search

International audience; Viral infections severely attack the physically frail elderly population, resulting in fatal drawbacks. The fact of having massive elderly population growth in Europe gives high priority to the detection of physical frailty and infectious diseases. This paper presents a safe, accurate, fast temperature detection system that could be integrated into homes or assisted living residences. The presented work aims to detect one of the symptoms of contagious diseases: elevated body temperature. In order to do so, we worked on recognizing eyes in thermal face images followed by scanning the detected eyes region for inner canthus temperature. Eyes detection was done by training four different sizes of You Only Look Once 5 th version (YOLOv5) object detection algorithm: nano, small, medium and large. A total of 4,255 thermal images were implemented for the training process after merging two different datasets and applying data augmentation techniques. Results show a similar mAP score (99.5%) for the different trained models. The large YOLOv5 model was the fastest, working at 115 FPS.

International audience; Emergency department is a key component of the health system where the management of crowding situations is crucial to the well-being of patients. This study proposes a new machine learning methodology and a queuing network model to measure and optimize crowding through a congestion indicator, which indicates a real-time level saturation.

Assessing the health profiles of populations is a crucial task to create a coherent healthcare offer. Emergency Departments (EDs) are at the core of the healthcare system and could benefit from this evaluation via an improved understanding of the healthcare needs of their population. This paper proposes a novel hierarchical agglomerative clustering algorithm based on multimorbidity analysis. The proposed approach constructs the clustering dendrogram by introducing new quality indicators based on the relative risk of co-occurrences of patient diagnoses. This algorithm enables the detection of multimorbidity patterns by merging similar patient profiles according to their common diagnoses. The multimorbidity approach has been applied to the data of the largest ED of the Aube Department (Eastern France) to cluster its patient visits. Among the 120,718 visits identified during a 24-month period, 16 clusters were identified, accounting for 94.8% of the visits, with the five most prevalent clusters representing 63.0% of them. The new quality indicators show a coherent and good clustering solution with a cluster membership of 1.81 based on a cluster compactness of 1.40 and a cluster separation of 0.77. Compared to the literature, the proposed approach is appropriate for the discovery of multimorbidity patterns and could help to develop better clustering algorithms for more diverse healthcare datasets.

International audience

Cardiovascular diseases are the major cause of death in the world. In this paper, we propose an automatic diagnostic method for cardiovascular diseases based on physiological signals. The problem is treated as a multi-label classification problem, where a set of labels is associated to each cardiovascular disease. Physiological signals are processed to extract features from them, according to each label or set of labels. The multi-label classification problem is then transformed to a set of two-class classification problems, that is solved using the belief functions theory. Association rules are also used to improve the quality of the classification. The simulation on the online medical database MIMIC III gives a Hamming loss of about 25% with the proposed method without the association rules. The Hamming loss reaches then 21% with the whole contribution.

International audience; This paper proposes a novel approach for predicting atrial fibrillation. The novelty of this approach remains in the reject option. Instead of classifying all instances, it rejects ambiguous observations by leaving them with no decision. In other words, three labels are considered: sick, healthy and unknown. Using physiological signals of patients, the proposed approach extracts several features from the signals in real-time. Then, it uses the features as sources of information in the belief functions framework. Due to the information sources, masses are assigned to the labels, and a risk level of being sick is computed. Prediction is assumed ambiguous if the risk level is within the rejection region and the state is unknown; otherwise, a decision is made. The approach is validated using the MIMIC III database. Due to the reject option, the prediction accuracy increases from 59% to around 78%.

International audience

International audience; La santé, le bien-être et l’autonomie d’une personne âgée dépendent pour partie de sa capacité à se mouvoir. Les évaluations cliniques de la mobilité sont pratiquées en consultation gériatrique au travers de tests standardisés. Cependant, une surveillance quotidienne peut fournir plus d’information s sur l’état fonctionnel d’une personne tout en préservant son autonomie à domicile. Cet article propose une nouvelle méthode d’une surveillance continue de la vitesse de la marche à travers un système radar Doppler placé à domicile en se basant sur l’analyse de la transformée en ondelettes. Plus de 70 ondelettes mères ont été testées sur des signaux de marche obtenus par le radar. Un système optoélectronique (Vicon) a été utilisé durant l’expérimentation afin de choisir la/les convenable(s) ondelette(s) mère(s). Le coefficient de corrélation de Pearson (ρ) a été calculé entre les signaux de vitesse de la marche obtenus par le système radar et ceux obtenus par le système Vicon. Nos résultats suggèrent l’utilisation de Daubechies5 et Symlet7 étant donné une valeur ρ de 0,86 et 0,85, respectivement et une valeur d’erreur quadratique inférieur à 0,05 m/s

International audience; Le suivi de trajectoires des personnes âgées est très important pour les aider le plus vite possible. Le présent article propose une méthode de suivi de trajectoires pour localiser des personnes âgées en temps réel dans des environnements fermés. Le modèle de mobilité est basé sur les modèles de Markov cachés. Ce modèle sert à estimer la trajectoire de la personne. Nous présentons dans cet article un modèle de mobilité en utilisant les modèles de Markov cachés. Ce modèle est combiné par un modèle d'observation. Ce dernier utilise les puissances des signaux WiFi reçus des points d'accès dans l'environnement. Les performances de la méthode proposée sont validées par des expérimentations sur des données réelles.

International audience; La santé, le bien-être et l’autonomie d’une personne âgée dépendent pour partie de sa capacité à se mouvoir. Les évaluations cliniques de la mobilité sont pratiquées en consultation gériatrique au travers de tests standardisés. Cependant, une surveillance quotidienne peut fournir plus d’informations sur l’état fonctionnel d’une personne tout en préservant son autonomie à domicile. Cet article propose une nouvelle méthode d’une surveillance continue de la vitesse de la marche à travers un système radar Doppler placé à domicile en se basant sur l’analyse de la transformée en ondelettes. Plus de 70 ondelettes mères ont été testées sur des signaux de marche obtenus par le radar. Un système optoélectronique (Vicon) a été utilisé durant l’expérimentation afin de choisir la/les convenable(s) ondelette(s) mère(s). Le coefficient de corrélation de Pearson (ρ) a été calculé entre les signaux de vitesse de la marche obtenus par le système radar et ceux obtenus par le système Vicon. Nos résultats suggèrent l’utilisation de Daubechies5 et Symlet7 étant donné une valeur ρ de 0,86 et 0,85, respectivement et une valeur d’erreur quadratique inférieur à 0,05 m/s.

Atrial Fibrillation is an irregularity in heart beats, increasing occurrence of heart failure, strokes, dementia and other serious diseases. The goal of this study is to predict Atrial Fibrillation using physiological signals of patients. Easily measured using non-invasive sensors, these signals include Heart Rate, Respiratory Rate, Peripheral Oxygen Saturation, Pulse and Central Venous Pressure. In order to predict the occurrence of Atrial Fibrillation, the proposed approach consists of extracting several parameters from signals and then constructing a decision rule that combines evidence based on the belief functions theory. The data used in this study are provided by the MIMIC III medical database. The classifier obtained is able to predict the Atrial Fibrillation with an accuracy of 70.49%, a sensitivity of 77.07% and a specificity of 63.9%.

International audience

International audience; Le syndrome de détresse respiratoire aigüe est une maladie qui perturbe le système respiratoire et peut entraîner la mort. Ce papier présente un modèle pour la prédiction de ce syndrome chez des patients ventilés en utilisant leurs signaux vitaux. Le modèle proposé est une approche individuelle consistant à détecter l’anormalité dans les signaux de chaque patient à comparer avec son état initial. Une décision est ensuite conçue par une fusion des décisions partielles des différents signaux analysés. Les résultats obtenus illustrent la capacité prédictive élevée de la méthode avec une sensibilité de 78% et une spécificité de 85%.

International audience; The aim of parametric classification is to predict the target class of a new sample, under the hypothesis of known fitted distribution. A major drawback of this approach is the uncertainty due to the imprecise modeling of the training samples. For this purpose, a belief functions framework is provided to take into account uncertainties. The proposed method investigates the belief functions theory to assign a confidence weight to each class for any new sample. This approach yields a confidence-weighted parametric classification method for multi-class problems. The performance of the proposed method is validated by experiments on real data for indoor localization and for facial image recognition.

This study aims to predict Acute Respiratory Distress Syndrome (ARDS) in hospitalized patients using their physiological signals such as heart rate, breathing rate, peripheral arterial oxygen saturation and mean airway blood pressure. A data fusion approach based on hypothesis testing was developed, and applied to mechanically ventilated subjects in the MIMIC II database. By combining the information extracted from the signals using an aggregation rule, we are able to enhance the sensitivity of the ARDS prediction process. As a result, we obtained a sensitivity of up to 85% for individual signals, reaching approximately 92% using the data fusion rule.