检索结果-内蒙古大学图书馆

2nd International Conference on Intelligent Systems and Pattern Recognition, ISPR 2022

作者： Feki, Ines Ammar, Sourour Kessentini, Yousri Digital Research Center of Sfax B.P. 275 Sakiet Ezzit Sfax3021 Tunisia SM@RTS: Laboratory of Signals systeMs aRtificial Intelligence and neTworkS Sfax Tunisia

ISBN: (纸本)9783031082764

Most of existing computer vision applications rely on models trained on supervised corpora, this is contradictory to what the world is seeing with the explosion of massive sets of unlabeled data. In the field of medical imaging for example, creating labels is extremely time-consuming because professionals should spend countless hours looking at images to manually annotate, segment, etc. Recently, several works are looking for solutions to the challenge of learning effective visual representations with no human supervision. In this work, we investigate the potential of using a self-supervised learning as a pretraining phase in improving the classification of radiographic images when the amount of available annotated data is small. To do that, we propose to use a self-supervised framework by pretraining a deep encoder with contrastive learning on a chest X-ray dataset using no labels at all, and then fine-tuning it using only few labeled data samples. We experimentally demonstrate that an unsupervised pretraining on unlabeled data is able to learn useful representation from Chest X-ray images, and only few labeled data samples are sufficient to reach the same accuracy of a supervised model learnt on the whole annotated dataset. © 2022, Springer Nature Switzerland AG.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

Transformer-based approach for joint handwriting and named entity recognition in historical documents

arXiv

引用

arXiv 2021年

作者： Rouhou, Ahmed Cheikh Dhiaf, Marwa Kessentini, Yousri Salem, Sinda Ben InstaDeep Digital Research Center of Sfax SM@RTS: Laboratory of Signals systeMs aRtificial Intelligence and neTworkS

The extraction of relevant information carried out by named entities in handwriting documents is still a challenging task. Unlike traditional information extraction approaches that usually face text transcription and named entity recognition as separate subsequent tasks, we propose in this paper an end-to-end transformer-based approach to jointly perform these two tasks. The proposed approach operates at the paragraph level, which brings two main benefits. First, it allows the model to avoid unrecoverable early errors due to line segmentation. Second, it allows the model to exploit larger bi-dimensional context information to identify the semantic categories, reaching a higher final prediction accuracy. We also explore different training scenarios to show their effect on the performance and we demonstrate that a two-stage learning strategy can make the model reach a higher final prediction accuracy. As far as we know, this work presents the first approach that adopts the transformer networks for named entity recognition in handwritten documents. We achieve the new state-of-the-art performance in the ICDAR 2017 Information Extraction competition using the Esposalles database, for the complete task, even though the proposed technique does not use any dictionaries, language modeling, or post-processing. Copyright © 2021, The Authors. All rights reserved.

关键词： Information retrieval

来源：评论

学校读者我要写书评

暂无评论

A Convoy of Ground Mobile Vehicles Protection using Cooperative UAVs-based System

A Convoy of Ground Mobile Vehicles Protection using Cooperat...

引用

International Symposium on Networks, Computers and Communications (ISNCC)

作者： Aicha Idriss Hentati Lamia Chaari Fourati Digital Research Center of Sfax (CRNS) Laboratory of Signals systeMs aRtificial Intelligence and neTworkS (SM@RTS)

The last decade has witnessed a growth related to the use of cooperative Unmanned Aerial Vehicles (UAVs) in several complex military and civilian applications. Among the cooperative-based UAV applications, this paper considers the protection of Convoy of Ground Mobile Vehicles (CGMV). The sketched problem is a canonical problem that requires coordinating a team of UAVs including their movements, tasks, and balancing resources. In this regard, we propose an efficient task allocation scheme. Besides that, the suggested approach is based on a hybrid strategy combining the circular path and the path ahead protection strategies involving two UAV sets. The first UAV set performs a circular path encircling all the vehicles in the convoy and protects them from the potential threats on the two lateral sides. The second UAV set explores the path ahead of the convoy. The conceived convoy protection scheme, called CGMVP-CUAVs, is validated by simulation using the NetLogo3D environment. Different experiments were conducted to analyze the efficiency of the proposed approach. Simulation results demonstrate the robustness of our strategy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Graph Neural Networks for End-to-End Information Extraction from Handwritten Documents

Graph Neural Networks for End-to-End Information Extraction ...

引用

IEEE Workshop on Applications of Computer Vision (WACV)

作者： Yessine Khanfir Marwa Dhiaf Emna Ghodhbani Ahmed Cheikh Rouhou Yousri Kessentini InstaDeep Digital Research Centre of Sfax Tunisia Systems Artificial Intelligence and Networks SM@RTS: Laboratory of Signals

Automating Information Extraction (IE) from handwritten documents is a challenging task due to the wide variety of handwriting styles, the presence of noise, and the lack of labeled data. In this work, we propose an end-toend encoder-decoder model, that incorporates transformers and Graph Convolutional Networks (GCN), to jointly perform Handwritten Text Recognition (HTR) and Named Entity Recognition (NER). The proposed architecture is mainly composed of two parts: a Sparse Graph Transformer Encoder (SGTE), to capture efficient representations of input text images while controlling the propagation of information through the model. The SGTE is followed by a transformer decoder enhanced with a GCN that combines the outputs of the last SGTE layer and the Multi-Head Attention (MHA) block to reinforce the alignment of visual features to characters and Named Entity (NE) tags, resulting in more robust learned representations. The proposed model shows promising results and achieves state-of-the-art performance on the IAM dataset, and in the ICDAR 2017 Information Extraction competition using the Esposalles database.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Machine Learning Models for Early Prediction of COVID-19 Infections Based on Clinical Signs

引用

SN Computer Science 2024年第1期5卷 1-10页

作者： Ben Ammar, Boulbaba Salem, Ali Ben Said, Mouna Ben Aouicha, Mohamed Data Engineering and Semantics Research Unit Faculty of Science University of Sfax Sfax Tunisia Digital Research Center of Sfax Ministry of Higher Education and Scientific Research Sfax Tunisia Signals systeMs aRtificial intelligence and neTworkS Laboratory (SM@RTS) Digital Research Center of Sfax Sfax Tunisia

Nowadays, the appearance of common symptoms, such as cough, fever, and loss of smell and taste, is the starting point of a battle against the coronavirus. The first standard method of COVID-19 infection assertion has become the RT-PCR test, which is however an uncomfortable solution for both patients and medical staff due to its high cost, timeliness, and false-negative result issue. This has raised the need for reliable automatic detection systems that aid in the early prediction of the COVID-19 infections with a lower cost. In this work, we aim at profiting from the Machine Learning (ML) advances to provide a reliable and low-cost COVID-19 prediction system. This system is based on the disease starting point, which is the patients’ clinical symptoms, that are still under-explored. We developed seven predictive models using traditional ML classification algorithms using a public dataset of obvious high-risk factors from patients’ clinical signs. The dataset has first undergone a pre-processing phase consisting of feature engineering and dataset resampling to deal with imbalanced dataset issue. Our best classification model is able to detect true positives and true negatives and weed out false positive and false negatives with an accuracy of 93%. © 2024, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： Clinical signs COVID-19 Early diagnosis Machine learning

来源：评论

学校读者我要写书评

暂无评论

MR Brain Image Segmentation Optimized by Using Ant Colony Algorithm

MR Brain Image Segmentation Optimized by Using Ant Colony Al...

引用

IEEE SSD International Multi-Conference on Systems, signals and Devices

作者： Dalenda Bouzidi Fahmi Ghozzi Ahmed Fakhfakh Laboratory of signals systems artificial intelligence and networks (SM@RTS) Digital Research Center of Sfax (CRNS) Tunisia

ISBN: (纸本)9781665414937;9781665430579

In the last decades, image processing has become an interesting task in the field of medical imaging. one of the most important medical imaging techniques used for diagnosis is the magnetic resonance image (MRI) which is considered to be a very useful medical tool for detecting the tumor progression of multiple sclerosis MS. The segmentation of brain tumors from MRI images is an effective technique. Several recent approaches for segmentation and classification of MRI sequences have been proposed for the automatic detection of MS outliers. This paper illustrates an algorithm based on the meta-heuristic “Ant Colonies Optimization ACO” for the segmentation of MRI images. We propose to apply the technique of ant colonies to estimate the segmentations of brain MRI images from a novel MR datasets of 30 MS patients, which were acquired with a 3T MR scanner with conventional sequences in order to optimize their overall rendering and to compare with consensual segmentation. The submitted algorithm is evaluated by using MATLAB GUI program.

关键词： Image segmentation Ant colony optimization Multiple sclerosis Statistical analysis Magnetic resonance imaging Tools Rendering (computer graphics)

来源：评论

学校读者我要写书评

暂无评论

A novel cooperative collision avoidance system for vehicular communication based on deep learning

引用

International Journal of Information Technology (Singapore) 2024年第3期16卷 1661-1675页

作者： Farhat, Wajdi Ben Rhaiem, Olfa Faiedh, Hassene Souani, Chokri Laboratory of Microelectronics and Instrumentation University of Monastir Monastir Tunisia Higher Institute of Applied Sciences and Technology Sousse University Sousse Tunisia Computer Science Department College of Science Northern Border University Arar Saudi Arabia Digital Research Center of SFax (CRNS) SM@RTS (Laboratory of Signals systeMs aRtificial Intelligence and neTworkS) Sfax Tunisia

Global road traffic injuries represent a significant safety challenge, with the highest fatality rates worldwide stemming from a combination of drivers' reckless behavior and the ever-expanding number of vehicles on the roads. In light of these factors, Autonomous Vehicles (AVs) have received a lot of attention as a technology that has the potential to revolutionize various industries and solve a lot of problems. The rise of the Internet of Vehicles (IoV) and Connected Vehicles (CVs) enables AVs to perceive and react swiftly in complex road scenarios, prioritizing safety. These vehicles prioritize individual safety in all situations, significantly enhancing road safety. Mobile Edge Computing (MEC) plays a crucial role thanks to its capacity for minimal latency and high bandwidth capabilities, enabling critical vehicular applications. In this paper, we introduce a cooperative collision avoidance system based on MEC, named 2CAS-MEC, that proactively identifies and locates road hazards. Our system is based on MEC servers positioned alongside roads, it consistently analyzes traffic and hazard data, issuing targeted alerts to nearby vehicles regarding potential risks. Particularly, our work introduces an advanced deep learning system that utilizes 5G, mobile edge computing, and cloud intelligence to prevent vehicle collisions. The experimental results show that our system outperforms the existing related works and exhibits the anticipated performance, effectively helping drivers in accident avoidance. © 2023, The Author(s), under exclusive licence to Bharati Vidyapeeth's Institute of Computer Applications and Management.

关键词： Cloud Connected and autonomous vehicles Deep learning IOV MEC Vehicle to-everything (V2X)

来源：评论

学校读者我要写书评

暂无评论

Implementation of a Photovoltaic Default Detection System using Deep Learning

Implementation of a Photovoltaic Default Detection System us...

引用

International Conference on Advanced Systems and Electric Technologies (IC_ASET)

作者： Amir Gargouri Fahmi Haddeji Laboratory of Signals systeMs aRtificial Intelligence and neTworkS ”SM@RTS” Sfax University Sfax Tunisia TDS Company Sfax Tunisia

ISBN: (数字)9798350384895

ISBN: (纸本)9798350384901

In this paper, we propose a very compact implementation of real-time default detection in photovoltaic (PV) based on deep learning. The goal is to propose an innovative solution for PV fault detection to enhance the reliability and performance of photovoltaic installations. This involves early detection of faults and proactive maintenance of equipment. The dataset was created using Roboflow software, which allows to efficiently edit and manipulate the images. Furthermore, we used the YOLOv8 (You Only Look Once) model for real-time detection, which it offers good accuracy while being fast. In fact, we implemented our solution on the Raspberry PI board to create a real-time and portable system. Such implementation results demonstrated the effectiveness of our system to detect PV default using various evaluation metrics, and its ability to operate in real-time.

关键词： Photovoltaic systems Deep learning YOLO Embedded systems Real-time systems Software Software reliability

来源：评论

学校读者我要写书评

暂无评论

A Scalable Intrusion Detection Approach for Industrial Internet of Things Based on Federated Learning and Attention Mechanism

A Scalable Intrusion Detection Approach for Industrial Inter...

引用

IEEE Symposium on Computers and Communications (ISCC)

作者： Mudhafar Nuaimi Lamia Chaari Fourati Bassem Ben Hamed Laboratory of Signals Systems Artificial Intelligence and Networks (SM@RTS) Digital Research Center of SFAX (CRNS) University of Sfax

The Industrial Internet of Things (IIoT) widespread adoption has prompted multiple breaches on IIoT devices by attackers. Thus this threatens the security of data for the end user. Recurrent neural networks (RNNs) have been used for intruder detection in IIoT because IIoT traffics are created consecutively, but they are unable to represent long traffic sequences and cannot be parallelized either. In order to solve these problems in this paper we have introduced the Attention technique which is applied to the encoding layer. We also used a federated learning (FL) approach to reduce the communication overhead of collecting data from each worker node and storing it in the cloud server in the case of a centralized model, thus preserving network scalability. With the use of the Edge-IIoT dataset, we test our suggested methodology. The outcomes of our FL experiment enable the system to scale.

关键词：

来源：评论

学校读者我要写书评

暂无评论

RFID IoT Architecture for smart Inventory Management: Security Integration

RFID IoT Architecture for Smart Inventory Management: Securi...

引用

International Wireless Communications and Mobile Computing Conference, IWCMC

作者： Amira Henaien Hadda Ben Elhadj Lamia Chaari Fourati Digital Research Center (CNRS) Elons SM@RTS (Laboratory of Signals systeMs aRtificial Intelligence and neTworkS) Sfax Tunisia

ISBN: (数字)9798350361261

ISBN: (纸本)9798350361278

This paper focuses on integrating security measures into the design of RFID-based IoT inventory systems, with a focus on RFID authentication, alarm messaging, and geofencing tracking. Our main contribution is a novel RFID-based IoT system featuring enhanced sensing capabilities and a robust authentication protocol. Our proposed solution includes a holistic architecture utilizing MQTT for efficient communication and cloud-based storage. Our proposed collaborative mutual authentication scheme is rigorously evaluated for security informally and formally using Proverif.

关键词： Wireless communication Protocols Authentication Collaboration Computer architecture Inventory management Sensors

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：