Université de technologie de Troyes open archive
Not a member yet
10722 research outputs found
Sort by
A Comparative Review of Deep-Learning Models for Deepfakes Detection
No full textInternational audienceThe development of generative AI has advanced significantly over the past few decades, enabling the creation of deepfake images that are increasingly difficult to distinguish from genuine photographs. The widespread availability of these models, which can be easily used, poses a substantial risk of spreading disinformation. Consequently, there is a pressing need for robust and reliable methods to identify images that have been created or altered using generative AI models. To address this need, a diverse range of methods and models have been developed. However, these approaches are often not exhaustively compared to one another, nor are they evaluated using a common reference dataset. Moreover, the majority of existing deepfake detection models rely on deep learning techniques, with numerous models available for feature extraction and detection, ranging from simple Convolutional Neural Networks (CNNs) to more advanced Vision Transformers (ViTs). To ensure the comparability and reproducibility of deepfake detection models, a standardized benchmark is urgently required to evaluate their performance across a large-scale, common dataset. Such a reference benchmark would facilitate the development of more effective and robust detection methods by providing insight into the strengths and weaknesses of existing AI-based approaches. In addition to establishing this benchmark, this paper also explores the challenges of combining different AI-based deepfake detection models and investigates various aggregation methods to further improve overall detection performance. A large-scale experiment involving almost 50 generative AI methods and over 40 deep learning-based feature extraction and detection models demonstrates the relevance of this study.</div
Appearance Defect Detection and Localisation using a Lightweight CNN-based Detector
No full textInternational audienceAutomatic visual inspection plays a crucial role in many industrial sectors to assist human operators. This paper studies the general problem of automatic detection of appearance defects with application on wheels surface quality control. An original method is proposed combining image processing and deep learning. This method exploits geometrical knowledge of the manufactured product, which allows splitting the image into homogeneous zones, over which a dedicated lightweight deep learning network is trained to detect and locate anomalies with the highest accuracy. Additionally, the present paper also addresses the issue of training a supervised AI architecture with a limited availability and imbalance dataset containing 100, 000 images but only 1, 000 with defects. We show on this dataset that the proposed lightweight CNN can achieve a high detection rate for low false-positive rates, which is the main goal for applications in an operational context.</div
Flexible hybrid nanostructured chemiresistive sensing platform for low-temperature trace hydrogen detection
No full textInternational audienc
First-principles prediction of half metallic-ferromagnetism in La0.25Sr0.75Sn0.4In0.25Ru0.35O3 and enhanced experimental electrical and magnetic behaviours
No full textInternational audienc
EMR-NG: New-Generation Solutions for Healthcare
No full textInternational audienceElectronic medical records (EMRs) play an essential role in the modernization of healthcare systems, providing a centralized platform to manage patient data. This solution helps to improve the quality of care while optimizing administrative processes. This article presents EMR-NG (Electronic Medical Record - New Generation), an innovative solution developed by a specialized team at ZAY Digital World, designed to facilitate care management through functionalities such as EMR creation, appointment management, and secure communication between healthcare professionals. Testing revealed an intuitive user interface and optimized workflows. Future enhancements include the integration of artificial intelligence for predictive care, interoperability with other hospital management systems, and enhanced security for sensitive data
Recent sizing, placement, and management techniques for individual and shared battery energy storage systems in residential areas: A review
No full textInternational audienceNowadays, the installation of renewable energy sources (RESs) is increasing rapidly in residential areas and is gaining growing interest. This is mainly due to the continuous increase in energy prices and the decrease of RESs installation costs for residential consumers. Among all these sources of green energy, solar and wind energy are the most dominants. The production of these two sources is strongly correlated to meteorological variables (e.g., solar radiation, wind) which are characterized by an intermittent nature. Therefore, the integration of these sources increases the concern of power quality in the electrical system. In order to mitigate this problem, battery energy storage systems (BESSs) were adopted as one of the viable solutions. Their fast response capability and geographical independence can improve energy efficiency and reliability in the power system. In addition to all these benefits, a BESS can further reduce consumer’s electricity bill by storing excess RESs’ generation and allowing price arbitrage. The high prices of BESSs and their high maintenance costs have manifested the installation of shared BESSs in residential communities. This type of installation is more financially advantageous to the consumers than the individual one by the fact that they share all the costs between them. However, for both types of installations, a proper planning must be done in order to determine the optimal characteristics of the BESS that will guarantee an optimal operation. This involves determining the optimal size, placement and operation of the BESS. Numerous methods have been proposed in literature to find the optimal planning and operation of individual and shared BESSs. This paper presents a comprehensive review in which recent planning and management approaches are analyzed. All studies included in this review are detailed and classified according to several criteria, such as the objective functions considered, the techniques used, and the limitations
A New Similarity-Based Classification Scheme of Drone Network Attacks
No full textInternational audienceThe Internet of Drones (IoD) paradigm is an emerging technology that has gained significant attention from the research community in the recent years. Drone networks are now widely utilized across various sectors and industries. However, like all emerging technologies, these networks face numerous challenges with security being the most critical one. The wireless nature of drone communication makes them vulnerable to a range of cyber attacks. Additionally, the limited resources of drones pose challenges for implementing effective security solutions. In this paper, we propose a novel classification that categorizes various attacks on drone networks based on their similarities. The primary objective is to highlight the commonalities between different cyber attacks on drone systems. This classification will serve as the foundation for developing future unified solutions capable of efficiently mitigating multiple attack types simultaneously
Bi-objective unrelated parallel machine scheduling problem under availability and energy constraints
No full textInternational audienceThe objective of this study is to address a scheduling issue arising in production workshops, specifically dealing with unrelated parallel machine scheduling integrating flexible and periodic maintenance interventions. Machines, while operational, consume varying amounts of energy based on their state: idle states consume less energy than active ones. Additionally, energy consumption is influenced by both the production job and the specific machine.Our goal is to minimize two functions: the first pertains to reducing production and maintenance earliness/tardiness, while the second focuses on minimizing energy consumption. To address this problem, a Mixed Integer Linear Program (MILP) is proposed. The ϵ—constrained method is employed for computing the Pareto front, and further adaptation involves applying the Multi-Objective Simulated Annealing algorithm (MOSA) to handle instances of substantial size.The proposed MILP model demonstrates the ability to accurately determine the Pareto front for up to 30 production jobs and 2 machines on literature benchmarks within a timeframe of less than 1 hour. Moreover, the computed metrics proves the effectiveness of the proposed MOSA
A dedicated acceptance sampling plan for quality inspection in textile industry
No full textInternational audienc
Efficient fabric anomaly detection: A transfer learning framework with expedited training times
No full textInternational audienceIn industrial quality control, anomaly detection plays a critical role in identifying defective products. However, because of the rarity and time-consuming nature of defect collection, training models often rely solely on defect-free samples. This necessitates the use of unsupervised anomaly-detection techniques trained exclusively on defect-free data. Alternatively, defect data can be synthesized to augment the dataset with defective samples. In the textile industry, expeditious model training is crucial to ensure a smooth production flow. Unfortunately, most unsupervised methods require extensive training time. This paper proposes a novel transfer learning approach designed to achieve training times in seconds while effectively adapting the model to the target domain of fabric anomaly detection. The key contributions of our method include significantly reduced training times, up to 10 times faster than current state-of-the-art methods, and comparable performance in anomaly detection, achieving results on par with state-of-the-art approaches on benchmark datasets (MVTEC Anomaly Detection, TILDA, AITEX and DAGM). Additionally, our approach improves inference times, ensuring expedited and efficient anomaly detection during production. The proposed method offers a practical and efficient solution for real-time industrial quality control