Université de technologie de Troyes open archive
Not a member yet
10722 research outputs found
Sort by
Taguchi DoE analysis and characterization of 17-4 PH stainless steel parts produced by material extrusion (MEX) process
No full textInternational audienc
Enhancing Wet and Dry Cough Classification with MFCC and Audio Augmentation
No full textInternational audienc
Cybersécurité : définitions, concepts, métiers
No full textInternational audienceCe livre sur la cybersécurité est structuré en trois parties : les notions clés de l'informatique et des données du cyber-espace ; les failles et les menaces exploitables par la cyber-criminalité ; les différents acteurs et métiers engagés dans la lutte pour la cybersécurité.Pour les étudiants, informaticiens ou curieux, c'est une porte d'entrée vers les fondamentaux de la cybersécurité qui posent plusieurs jalons essentiels à la compréhension des enjeux du numérique : des schémas, des QCM et leurs corrigés sur un sujet qui concerne tout un chacun
Enhanced Vehicle Detection Mechanism for Traffic Management in Smart Cities
No full textInternational audienceNowadays Road traffic is a major issue in developing and under-developing countries. With the rampant increase in traffic, society faces major traffic threats including life threats and environmental threats, thus traffic management is a gruesome problem to address. The consequences of poor traffic management include road accidents, jamming traffic, pollution, and many more that can be life-threatening. Living in the twenty-first century with the emergence of technology and the applicability of smart cities provides a perfect solution to curb traffic issues. Keeping in view the deadlock and congestion in traffic, this work will provides solution by indigently detecting and prioritizing vehicles and non-vehicles. The research involves the implementation and comparison of two states of art algorithms Aggregated channel feature and a Point Tracker. Further, the algorithms are enhanced by improving traffic management in terms of identifying the category of transport, prioritizing the traffic which contains vehicles and non-vehicles on basis of the size of vehicle, type of the vehicle, and emergency providing priority to resolve the deadlock. Further, the proposed enhanced point tracker algorithm includes emergency detection in case of an accident and provides an alternative route to neighboring vehicles and non-vehicles. Enhanced ACF has detected a true positive rate of 80%, 89%, has detected true positive rates of 69%, and 79% having non-vehicle detected with assigned priority. Enhanced point tracker has detected true positive rate of 88%, 94%, and 86% having vehicles, non-vehicles, and assigned priority
Optimizing Acceptance Sampling for Enhanced Quality Control: A Data-Driven Approach with Criticality Assessment
No full textInternational audienceQuality control is a critical aspect of manufacturing and production processes. This paper focuses on optimizing acceptance sampling procedures to enhance control efficiency while acting on the criticality of the products. We introduce a novel Key Performance Indicator (KPI) for product criticality based on data analysis, enabling better decision-making in quality control. This research aims to improve the effectiveness of acceptance sampling and quality control, offering valuable insights for manufacturing and production industries
Joint optimization of inventory replenishment and rationing policies for an omnichannel store with both in-store and online demands
No full textInternational audienc
Toward Universal Detector for Synthesized Images by Estimating Generative AI Models
No full textInternational audienceOne of the vulnerabilities in discriminators for AI-generated images is that the classification accuracy degrades when dealing with images generated using methods other than those they were trained on. As a countermeasure, in this study, we propose an image generation method estimator. The process of discrimination involves the input of an image to the estimator, which estimates the method used for its generation. Subsequently, a specialized fake image discriminator tailored to the estimated image generation method is used to identify the authenticity of the image. The activation functions are also considered according to the estimation results and analyzed for those discriminators. Discrimination scores are weighted and aggregated according to the estimation results, and the final decision is output. Our experimental results showed that the estimator achieved a classification accuracy of approximately 90% for 18 types of AI-generated images. Furthermore, by selecting the top two estimations in order of confidence, the accuracy increased to around 98%
Path Planning in UAV-Assisted Wireless Networks: A Comprehensive Survey and Open Research Issues
No full textInternational audienceUAV-assisted wireless networks are becoming more and more used and are invading many fields thanks to their performance and efficiency. However, there are still some challenges that need to be addressed before this technology can be widely adopted. One of the main issues is UAV path planning. This task is challenging due to various factors such as data collection, energy consumption, limited battery life, and dynamic changes in the environment. Efficient path planning algorithms are crucial to ensuring safe and efficient UAV operations, minimizing collision risks, and maximizing mission success. To give a complete and clear view of recent papers dealing with this crucial trajectory tracing problem, this survey aims to present a collection of work carried out in this line of research, and for ease of convenience, we have classified existing solutions according to the optimization method selected: heuristic, genetic, machine learning and game theory. Our analysis and qualitative comparison of the current literature on UAV path planning, unveil open research challenges in this field. These challenges serve as a roadmap for future research efforts in the deployment of UAV-assisted wireless networks and steer the exploration of innovative solutions
Dual model knowledge distillation for industrial anomaly detection
No full textInternational audienceUnsupervised texture anomaly detection has been a concerning topic in a vast number of industrial processes. Patterned textures inspection, particularly in the context of fabric defect detection, is indeed a widely encountered use case. This task involves handling a diverse spectrum of colors and textile types, encompassing a wide range of fabrics. Given the extensive variability in colors, textures, and defect types, fabric defect detection poses a complex and challenging problem in the field of patterned textures inspection. In this article, we propose a knowledge distillation-based approach tailored specifically for addressing the challenge of unsupervised anomaly detection in textures resembling fabrics. Our method aims to redefine the recently introduced reverse distillation approach, which advocates for an encoder-decoder design to mitigate classifier bias and to prevent the student from reconstructing anomalies. In this study, we present a new reverse distillation technique for the specific task of fabric defect detection. Our approach involves a meticulous design selection that strategically highlights high-level features. To demonstrate the capabilities of our approach both in terms of performance and inference speed, we conducted a series of experiments on multiple texture datasets, including Machine Vision Technology Development Corporation (MVTEC) anomaly detection, Asociación de Investigación de la Industria Textil (AITEX), and TILDA, alongside conducting experiments on a dataset acquired from a textile manufacturing facility. The main contributions of this paper are the following: a robust texture anomaly detector utilizing a reverse knowledge-distillation technique suitable for both anomaly detection and domain generalization and a novel dataset encompassing a diverse range of fabrics and defects
