Scientific Publications of the University of Toulouse II Le Mirail
Not a member yet
92205 research outputs found
Sort by
A Noninvasive Framework for Heart Function Assessment by Multitask Learning
No full textInternational audienceAccurate assessment of cardiac function is vital for preventing and managing cardiovascular diseases (CVDs). Recent advancements in machine learning, especially convolutional neural networks (CNNs) and multitask learning (MTL), have improved the precision of echocardiogram evaluations. However, existing methods often overlook the intrinsic relationships among ejection fraction (EF), end-diastolic volume (EDV), and end-systolic volume (ESV), which are essential for accurate assessments. We propose a noninvasive framework for heart function assessment (FHFA) using MTL that utilizes a 3-D CNN to extract key spatiotemporal features from echocardiogram videos. By employing an MTL strategy and weight distribution mechanism, this framework enhances the accuracy of EF predictions and provides a comprehensive assessment of cardiac structure and function. This approach ensures that the model effectively integrates auxiliary task information while focusing on the primary task, resulting in a more precise analysis of cardiac function. The experimental results on the Echonet-Dynamic dataset demonstrate that our method achieves an average absolute error of 3.89, a root-mean-square error (RMSE) of 5.13, and an R2 value of 0.82, outperforming existing methods. Future work will focus on automatic weight optimization, model compression, and improving computational efficiency for broader clinical applications
Le gouvernement d’Amanieu II, archevêque d’Auch (1261-1318)
No full textInternational audienceThis paper, which stems from doctoral studies in medieval history, is devoted to the governmental style of a 13th-century archbishop in Gascony who enjoyed an impressive degree of longevity. Based on original textual sources, it portrays his style of governance, which combined lordship power with leadership of the Christian people. The case study is examined through various topics, such as the archbishop's territory, including the control of tithes, and the church's dominion in relation to secular lordship. The analysis aims to depict Amanieu's government style and the human relations that formed the basis of his leadership over the clergy and his role in guiding the Christian people.Issu des recherches d'un doctorat d'histoire médiévale, cet article étudie le mode de gouvernement d'un archevêque gascon du XIIIe siècle, à la longévité impressionnante. S'appuyant sur l'examen de sources originelles, il tente de cerner son art de gouverner, mêlant domination seigneuriale et encadrement du peuple chrétien. L'analyse aborde deux angles thématiques : le territoire de l'archevêque avec le contrôle des dîmes, et la nature du dominium de l'Église d'Auch face aux pouvoirs laïcs. Cette analyse tentera de déceler le mode de gouvernement d'Amanieu, et les ressorts humains qui animent sa direction du clergé et son encadrement du peuple chrétien
Impact de l'ajustement dynamique de la difficulté du jeu sur le sentiment de flow
No full textInternational audienceDynamic difficulty adjustment (DDA) is a key mechanism for maintaining player engagement in video games. This study compares two types of DDA: one based on player performance, and the other on pupil size, a physiological indicator reflecting arousal and cognitive load. Forty-four participants played two sessions of Tetris using these two adjustment methods. Flow experience was measured using a standardized scale and an innovative temporal graphtracing method. Although no significant difference was observed on average in flow scores, exploratory results show that pupil-based adjustment leads to more frequent changes in game difficulty and a generally higher effective level of challenge, without degrading performance or the players' flow experience. These results highlight the potential of difficulty adjustments based on the player's physiological state and open up new perspectives for affective game design
Joy Harjo's biocentric poetry: sharing a conversation with all-our-relations “inside the raw cortex of songs”
No full textInternational audienc
Modélisation à temps continu du lien entre la centration du soi et le bonheur
No full textInternational audienc
Saudi Arabian-French Joint Archaeological Project in al-Ukhdūd, Province of Najrān, Saudi Arabia. Report of the first field season — April 2025
No full textThe first season of the Saudi Arabian–French Joint Archaeological Project at al-Ukhdūd (Najrān Province, Saudi Arabia) took place from April 4 to 29, 2025, marking the beginning of a five-year collaborative program. Building on decades of Saudi investigations, this campaign aimed to reassess the site’s urban development, chronology, and cultural significance through an integrated strategy combining survey, excavation, and digital documentation. A systematic pedestrian survey mapped nearly 60% of the 201-hectare site, recording over 550 features including residential units, workshops, funerary areas, religious structures, and defensive installations. Extensive drone-based photogrammetry and RTK-GPS control produced high-resolution orthophotographs and digital surface models covering 117 hectares, complemented by mobile LiDAR scans of structures and select artefacts. Excavations focused on Building A43 (UK004) within the fortified residential sector and a deep test pit (UK004-TP1), revealing stratified architectural phases from the late pre-Islamic to early Islamic periods. Radiocarbon dates situated key occupation phases between the 3rd century BCE and 7th century CE. These results confirm al-Ukhdūd’s role as a major urban centre in South Arabia that bridged pre-Islamic and Islamic periods, providing new insights into its architectural and historical trajectories
El mundo de los traductores de Toledo: perspectivas sociales y culturales del círculo de los canónigos de la catedral (1150-1250)
No full textInternational audienc
Habiter en vagabond : l'absence de maison dans Lonesome Traveler de Jack Kerouac
No full textInternational audienc
Detecting Anomalies Using Graph Neural Networks: A Review
No full textInternational audienceAnomaly detection is the process of identifying unusual behaviors in systems. In this wide-ranging field, graph neural networks (GNNs) are highly effective compared to the other proposed approaches in the literature. This article summarizes the representative GNN-based methods for anomaly detection and proposes a novel taxonomy based on how these methods predict anomalies.La détection d’anomalies vise à identifier des comportements atypiques au sein des systèmes complexes. Parmi les différentes approches développées dans ce domaine, les réseaux de neurones graphiques (GNN) se distinguent par leur efficacité. Dans cet article, nous proposons une revue des méthodes fondées sur les GNN pour la détection d’anomalies, et introduisons une nouvelle taxonomie, construite autour des mécanismes de prédiction d’anomalies utilisés
