RESUMEN
In recent years, Federated Learning (FL) has gained traction as a privacy-centric approach in medical imaging. This study explores the challenges posed by data heterogeneity on FL algorithms, using the COVIDx CXR-3 dataset as a case study. We contrast the performance of the Federated Averaging (FedAvg) algorithm on non-identically and independently distributed (non-IID) data against identically and independently distributed (IID) data. Our findings reveal a notable performance decline with increased data heterogeneity, emphasizing the need for innovative strategies to enhance FL in diverse environments. This research contributes to the practical implementation of FL, extending beyond theoretical concepts and addressing the nuances in medical imaging applications. This research uncovers the inherent challenges in FL due to data diversity. It sets the stage for future advancements in FL strategies to effectively manage data heterogeneity, especially in sensitive fields like healthcare.
Asunto(s)
Algoritmos , Diagnóstico por Imagen , Humanos , Diagnóstico por Imagen/métodos , COVID-19/epidemiología , COVID-19/diagnóstico por imagen , Aprendizaje Automático , SARS-CoV-2/aislamiento & purificaciónRESUMEN
Semantic segmentation of Remote Sensing (RS) images involves the classification of each pixel in a satellite image into distinct and non-overlapping regions or segments. This task is crucial in various domains, including land cover classification, autonomous driving, and scene understanding. While deep learning has shown promising results, there is limited research that specifically addresses the challenge of processing fine details in RS images while also considering the high computational demands. To tackle this issue, we propose a novel approach that combines convolutional and transformer architectures. Our design incorporates convolutional layers with a low receptive field to generate fine-grained feature maps for small objects in very high-resolution images. On the other hand, transformer blocks are utilized to capture contextual information from the input. By leveraging convolution and self-attention in this manner, we reduce the need for extensive downsampling and enable the network to work with full-resolution features, which is particularly beneficial for handling small objects. Additionally, our approach eliminates the requirement for vast datasets, which is often necessary for purely transformer-based networks. In our experimental results, we demonstrate the effectiveness of our method in generating local and contextual features using convolutional and transformer layers, respectively. Our approach achieves a mean dice score of 80.41%, outperforming other well-known techniques such as UNet, Fully-Connected Network (FCN), Pyramid Scene Parsing Network (PSP Net), and the recent Convolutional vision Transformer (CvT) model, which achieved mean dice scores of 78.57%, 74.57%, 73.45%, and 62.97% respectively, under the same training conditions and using the same training dataset.
RESUMEN
Wireless sensor networks (WSNs), constrained by limited resources, demand routing strategies that prioritize energy efficiency. The tactic of cooperative routing, which leverages the broadcast nature of wireless channels, has garnered attention for its capability to amplify routing efficacy. This manuscript introduces a power-conscious routing approach, tailored for resource-restricted WSNs. By exploiting cooperative communications, we introduce an innovative relay node selection technique within clustered networks, aiming to curtail energy usage while safeguarding data dependability. This inventive methodology has been amalgamated into the Routing Protocol for Low-Power and Lossy Networks (RPL), giving rise to the cooperative and efficient routing protocol (CERP). The devised CERP protocol pinpoints and selects the most efficacious relay node, ensuring that packet transmission is both energy-minimal and reliable. Performance evaluations were executed to substantiate the proposed strategy, and its practicality was examined using an Arduino-based sensor node and the Contiki operating system in real-world scenarios. The outcomes affirm the efficacy of the proposed strategy, outshining the standard RPL concerning reliability and energy conservation, enhancing RPL reliability by 10% and energy savings by 18%. This paper is posited to contribute to the evolution of power-conscious routing strategies for WSNs, crucial for prolonging sensor node battery longevity while sustaining dependable communication.
RESUMEN
Since the release of ChatGPT, numerous studies have highlighted the remarkable performance of ChatGPT, which often rivals or even surpasses human capabilities in various tasks and domains. However, this paper presents a contrasting perspective by demonstrating an instance where human performance excels in typical tasks suited for ChatGPT, specifically in the domain of computer programming. We utilize the IEEExtreme Challenge competition as a benchmark-a prestigious, annual international programming contest encompassing a wide range of problems with different complexities. To conduct a thorough evaluation, we selected and executed a diverse set of 102 challenges, drawn from five distinct IEEExtreme editions, using three major programming languages: Python, Java, and C++. Our empirical analysis provides evidence that contrary to popular belief, human programmers maintain a competitive edge over ChatGPT in certain aspects of problem-solving within the programming context. In fact, we found that the average score obtained by ChatGPT on the set of IEEExtreme programming problems is 3.9 to 5.8 times lower than the average human score, depending on the programming language. This paper elaborates on these findings, offering critical insights into the limitations and potential areas of improvement for AI-based language models like ChatGPT.
RESUMEN
The Vision Transformer (ViT) architecture has been remarkably successful in image restoration. For a while, Convolutional Neural Networks (CNN) predominated in most computer vision tasks. Now, both CNN and ViT are efficient approaches that demonstrate powerful capabilities to restore a better version of an image given in a low-quality format. In this study, the efficiency of ViT in image restoration is studied extensively. The ViT architectures are classified for every task of image restoration. Seven image restoration tasks are considered: Image Super-Resolution, Image Denoising, General Image Enhancement, JPEG Compression Artifact Reduction, Image Deblurring, Removing Adverse Weather Conditions, and Image Dehazing. The outcomes, the advantages, the limitations, and the possible areas for future research are detailed. Overall, it is noted that incorporating ViT in the new architectures for image restoration is becoming a rule. This is due to some advantages compared to CNN, such as better efficiency, especially when more data are fed to the network, robustness in feature extraction, and a better feature learning approach that sees better the variances and characteristics of the input. Nevertheless, some drawbacks exist, such as the need for more data to show the benefits of ViT over CNN, the increased computational cost due to the complexity of the self-attention block, a more challenging training process, and the lack of interpretability. These drawbacks represent the future research direction that should be targeted to increase the efficiency of ViT in the image restoration domain.
RESUMEN
Video streaming-based real-time vehicle identification and license plate recognition systems are challenging to design and deploy in terms of real-time processing on edge, dealing with low image resolution, high noise, and identification. This paper addresses these issues by introducing a novel multi-stage, real-time, deep learning-based vehicle identification and license plate recognition system. The system is based on a set of algorithms that efficiently integrate two object detectors, an image classifier, and a multi-object tracker to recognize car models and license plates. The information redundancy of Saudi license plates' Arabic and English characters is leveraged to boost the license plate recognition accuracy while satisfying real-time inference performance. The system optimally achieves real-time performance on edge GPU devices and maximizes models' accuracy by taking advantage of the temporally redundant information of the video stream's frames. The edge device sends a notification of the detected vehicle and its license plate only once to the cloud after completing the processing. The system was experimentally evaluated on vehicles and license plates in real-world unconstrained environments at several parking entrance gates. It achieves 17.1 FPS on a Jetson Xavier AGX edge device with no delay. The comparison between the accuracy on the videos and on static images extracted from them shows that the processing of video streams using this proposed system enhances the relative accuracy of the car model and license plate recognition by 13% and 40%, respectively. This research work has won two awards in 2021 and 2022.
RESUMEN
Extensive research has been conducted on healthcare technology and service advancements during the last decade. The Internet of Medical Things (IoMT) has demonstrated the ability to connect various medical apparatus, sensors, and healthcare specialists to ensure the best medical treatment in a distant location. Patient safety has improved, healthcare prices have decreased dramatically, healthcare services have become more approachable, and the operational efficiency of the healthcare industry has increased. This research paper offers a recent review of current and future healthcare applications, security, market trends, and IoMT-based technology implementation. This research paper analyses the advancement of IoMT implementation in addressing various healthcare concerns from the perspectives of enabling technologies, healthcare applications, and services. The potential obstacles and issues of the IoMT system are also discussed. Finally, the survey includes a comprehensive overview of different disciplines of IoMT to empower future researchers who are eager to work on and make advances in the field to obtain a better understanding of the domain.
RESUMEN
[This corrects the article DOI: 10.1007/s40747-022-00767-w.].
RESUMEN
Autonomous robots require control tuning to optimize their performance, such as optimal trajectory tracking. Controllers, such as the Proportional-Integral-Derivative (PID) controller, which are commonly used in robots, are usually tuned by a cumbersome manual process or offline data-driven methods. Both approaches must be repeated if the system configuration changes or becomes exposed to new environmental conditions. In this work, we propose a novel algorithm that can perform online optimal control tuning (OCTUNE) of a discrete linear time-invariant (LTI) controller in a classical feedback system without the knowledge of the plant dynamics. The OCTUNE algorithm uses the backpropagation optimization technique to optimize the controller parameters. Furthermore, convergence guarantees are derived using the Lyapunov stability theory to ensure stable iterative tuning using real-time data. We validate the algorithm in realistic simulations of a quadcopter model with PID controllers using the known Gazebo simulator and a real quadcopter platform. Simulations and actual experiment results show that OCTUNE can be effectively used to automatically tune the UAV PID controllers in real-time, with guaranteed convergence. Finally, we provide an open-source implementation of the OCTUNE algorithm, which can be adapted for different applications.
RESUMEN
Pristine and trustworthy data are required for efficient computer modelling for medical decision-making, yet data in medical care is frequently missing. As a result, missing values may occur not just in training data but also in testing data that might contain a single undiagnosed episode or a participant. This study evaluates different imputation and regression procedures identified based on regressor performance and computational expense to fix the issues of missing values in both training and testing datasets. In the context of healthcare, several procedures are introduced for dealing with missing values. However, there is still a discussion concerning which imputation strategies are better in specific cases. This research proposes an ensemble imputation model that is educated to use a combination of simple mean imputation, k-nearest neighbour imputation, and iterative imputation methods, and then leverages them in a manner where the ideal imputation strategy is opted among them based on attribute correlations on missing value features. We introduce a unique Ensemble Strategy for Missing Value to analyse healthcare data with considerable missing values to identify unbiased and accurate prediction statistical modelling. The performance metrics have been generated using the eXtreme gradient boosting regressor, random forest regressor, and support vector regressor. The current study uses real-world healthcare data to conduct experiments and simulations of data with varying feature-wise missing frequencies indicating that the proposed technique surpasses standard missing value imputation approaches as well as the approach of dropping records holding missing values in terms of accuracy.
RESUMEN
For many years, mental health has been hidden behind a veil of shame and prejudice. In 2017, studies claimed that 10.7% of the global population suffered from mental health disorders. Recently, people started seeking relaxing treatment through technology, which enhanced and expanded mental health care, especially during the COVID-19 pandemic, where the use of mental health forums, websites, and applications has increased by 95%. However, these solutions still have many limits, as existing mental health technologies are not meant for everyone. In this work, an up-to-date literature review on state-of-the-art of mental health and healthcare solutions is provided. Then, we focus on Arab-speaking patients and propose an intelligent tool for mental health intent recognition. The proposed system uses the concepts of intent recognition to make mental health diagnoses based on a bidirectional encoder representations from transformers (BERT) model and the International Neuropsychiatric Interview (MINI). Experiments are conducted using a dataset collected at the Military Hospital of Tunis in Tunisia. Results show excellent performance of the proposed system (the accuracy is over 92%, the precision, recall, and F1 scores are over 94%) in mental health patient diagnosis for five aspects (depression, suicidality, panic disorder, social phobia, and adjustment disorder). In addition, the tool was tested and evaluated by medical staff at the Military Hospital of Tunis, who found it very interesting to help decision-making and prioritizing patient appointment scheduling, especially with a high number of treated patients every day.
Asunto(s)
COVID-19 , Salud Mental , Humanos , Pandemias , Escalas de Valoración Psiquiátrica , SARS-CoV-2RESUMEN
The whole world is facing a health crisis, that is unique in its kind, due to the COVID-19 pandemic. As the coronavirus continues spreading, researchers are concerned by providing or help provide solutions to save lives and to stop the pandemic outbreak. Among others, artificial intelligence (AI) has been adapted to address the challenges caused by pandemic. In this article, we design a deep learning system to extract features and detect COVID-19 from chest X-ray images. Three powerful networks, namely ResNet50, InceptionV3, and VGG16, have been fine-tuned on an enhanced dataset, which was constructed by collecting COVID-19 and normal chest X-ray images from different public databases. We applied data augmentation techniques to artificially generate a large number of chest X-ray images: Random Rotation with an angle between - 10 and 10 degrees, random noise, and horizontal flips. Experimental results are encouraging: the proposed models reached an accuracy of 97.20 % for Resnet50, 98.10 % for InceptionV3, and 98.30 % for VGG16 in classifying chest X-ray images as Normal or COVID-19. The results show that transfer learning is proven to be effective, showing strong performance and easy-to-deploy COVID-19 detection methods. This enables automatizing the process of analyzing X-ray images with high accuracy and it can also be used in cases where the materials and RT-PCR tests are limited.
RESUMEN
Unmanned aerial systems (UAVs) are dramatically evolving and promoting several civil applications. However, they are still prone to many security issues that threaten public safety. Security becomes even more challenging when they are connected to the Internet as their data stream is exposed to attacks. Unmanned traffic management (UTM) represents one of the most important topics for small unmanned aerial systems for beyond-line-of-sight operations in controlled low-altitude airspace. However, without securing the flight path exchanges between drones and ground stations or control centers, serious security threats may lead to disastrous situations. For example, a predefined flight path could be easily altered to make the drone perform illegal operations. Motivated by these facts, this paper discusses the security issues for UTM's components and addresses the security requirements for such systems. Moreover, we propose UTM-Chain, a lightweight blockchain-based security solution using hyperledger fabric for UTM of low-altitude UAVs which fits the computational and storage resources limitations of UAVs. Moreover, UTM-Chain provides secure and unalterable traffic data between the UAVs and their ground control stations. The performance of the proposed system related to transaction latency and resource utilization is analyzed by using cAdvisor. Finally, the analysis of security aspects demonstrates that the proposed UTM-Chain scheme is feasible and extensible for the secure sharing of UAV data.
RESUMEN
Unmanned Aerial Vehicles (UAVs) have been very effective in collecting aerial images data for various Internet-of-Things (IoT)/smart cities applications such as search and rescue, surveillance, vehicle detection, counting, intelligent transportation systems, to name a few. However, the real-time processing of collected data on edge in the context of the Internet-of-Drones remains an open challenge because UAVs have limited energy capabilities, while computer vision techniquesconsume excessive energy and require abundant resources. This fact is even more critical when deep learning algorithms, such as convolutional neural networks (CNNs), are used for classification and detection. In this paper, we first propose a system architecture of computation offloading for Internet-connected drones. Then, we conduct a comprehensive experimental study to evaluate the performance in terms of energy, bandwidth, and delay of the cloud computation offloading approach versus the edge computing approach of deep learning applications in the context of UAVs. In particular, we investigate the tradeoff between the communication cost and the computation of the two candidate approaches experimentally. The main results demonstrate that the computation offloading approach allows us to provide much higher throughput (i.e., frames per second) as compared to the edge computing approach, despite the larger communication delays.