Energy-efficient routing protocols for UWSNs: A comprehensive review of taxonomy, challenges, opportunities, future research directions, and machine learning perspectives

Journal of King Saud University - Computer and Information Sciences, Год журнала: 2024, Номер 36(7), С. 102128 - 102128

Опубликована: Июль 24, 2024

Underwater Wireless Sensor Networks (UWSNs) are essential for a number of environmental and oceanographic monitoring applications. However, they face different more complex challenges than terrestrial wireless sensor networks (TWSNs). The main faced by UWSNs limited include high propagation delays, poor bandwidth, low throughput, energy consumption. Replacing batteries in such becomes extremely difficult as usually deployed remote areas where human interaction is possible. unbalanced inefficient usage various network nodes poses another issue, it may reduce the applicability feasibility network. Therefore, proposing Energy-Efficient Routing Protocols (E-ER-Ps) crucial to improve performance lifespan these networks. Due mentioned earlier, this research presents an extensive analysis several E-ER-Ps intended UWSNs. We compare contemporary approaches that use machine learning (ML) with conventional protocols, ML-based have shown significant potential resolving intricate This paper aims present critical review from prospects To better comprehend structure uses we provide innovative taxonomy their classification. While protocols evaluated flexibility, predictive power, overall efficiency advancements, traditional based on routing tactics energy-efficiency improvements. A thorough comparative highlights advantages, disadvantages, possible protocols. Furthermore, ML's function, incorporating intelligent adaptive approaches, presented, highlighting technology's completely alter UWSN management. formulate implement UWSNs, article concludes obstacles, including need real-time resilience alters, pre-existing infrastructures. development hybrid combine methodologies, design can adapt dynamically changing circumstances underwater habitats highlighted future objectives. provides foundation advancements field presenting comprehensive overview state-of-the-art E-ER-Ps.

Язык: Английский

---

Advances in Thermal Imaging: A Convolutional Neural Network Approach for Improved Breast Cancer Diagnosis

Опубликована: Март 15, 2024

This study explores the application of thermal imaging in breast cancer diagnostics, presenting a novel methodology that integrates pre-processing techniques and Convolutional Neural Networks (CNN) for classification images. A comprehensive dataset from DMR Database is used, comprising balanced samples normal cancerous The images are augmented, standardized, enhanced through data augmentation, resizing, filtering, with crucial features extracted via Histogram Oriented Gradients (HOG). CNN model, specifically constructed this study, then trained tested on these processed approach's effectiveness benchmarked against other classifiers, displaying promising results accuracy rates ranging 95.7% to 98.5%. Future work suggests exploring fusion traditional advanced modalities, expanding datasets, utilizing pre-trained models improved diagnostic precision. research signifies step toward development real-time, efficient tools cancer, highlighting potential impact patient outcomes broader medical field.

Язык: Английский

---

Advancing diabetic retinopathy diagnosis with fundus imaging: A comprehensive survey of computer-aided detection, grading and classification methods

Global Transitions, Год журнала: 2024, Номер 6, С. 93 - 112

Опубликована: Янв. 1, 2024

The incidence of diabetic retinopathy globally calls for advanced and more universally applicable computer-aided diagnosis (CAD) systems. This survey paper explores the current state vision-based CAD techniques detection classification retinopathy, a diabetes-induced eye disorder that can lead to severe visual impairment or blindness. Characterized by variety manifestations including microaneurysms, exudates, hemorrhages, macular detachment, presents substantial challenges automated detection. is primarily due heterogeneity retinal fundus images, which display diverse spatiotextural features intricate vascular structures. Our exhaustive review indicates most existing methodologies predominantly concentrate on isolated types, employing localized feature analysis classification. Such specificity often results in limited accuracy generalizability, restricting practical real-world application. Furthermore, contemporary leading methods generally focus single characteristics, necessitating patients undergo multiple procedures, thereby increasing time, costs, possibly intensifying complexities. To overcome these obstacles, we propose adoption multi-trait-driven solutions. Utilizing potent capabilities deep learning, solutions could employ high-dimensional, multi-cue sensitive extraction ensemble learning approach designed improve generalizability dependability systems, offering holistic solution capable effectively managing retinopathy. study highlights need fundamental transformation motivating further research towards robust, multi-modal enhance detection, classification, grading this widespread ailment.

Язык: Английский

---

Leveraging Transfer Learning for Efficient Diagnosis of COPD Using CXR Images and Explainable AI Techniques

INTELIGENCIA ARTIFICIAL, Год журнала: 2024, Номер 27(74), С. 133 - 151

Опубликована: Июнь 12, 2024

Chronic Obstructive Pulmonary Disease (COPD) is a predominant global health concern, ranking third in mortality rates, yet frequently remains undiagnosed until its advanced stages. Given prevalence, the need for innovative and widely accessible diagnostic tools has never been more paramount. While spirometry tests serve as conventional benchmarks, their reach limited, especially regions with constrained medical resources. The presented research harnesses deep learning algorithms to facilitate early-stage COPD detection, specifically targeting Chest X-rays (CXRs). clinically annotated VinDR-CXR dataset provides primary foundation model training, complemented by incorporating ChestX-ray14 initial pre-training. Such dual-dataset strategy augments generalization adaptability. Among several explored Convolutional Neural Network (CNN) architectures, Xception emerges frontrunner. Through transfer methodologies, this produces noteworthy recall rate of 98.2%, markedly surpassing metrics ResNet50 model. Recognizing imperative transparency AI applications imaging, integrates essential explainability approaches viz: Gradient Class Activation Mapping (Grad-CAM) SHapley Additive exPlanations (SHAP). These techniques elucidate AI’s decision-making process, offering invaluable visual analytical insights fostering trust among professionals. In essence, study not only underscores potential integrating imaging detection but also accentuates pivotal role AI-driven interventions.

Язык: Английский

---

Enhanced diabetic retinopathy detection and classification using fundus images with ResNet50 and CLAHE-GAN

Indonesian Journal of Electrical Engineering and Computer Science, Год журнала: 2024, Номер 35(1), С. 366 - 366

Опубликована: Май 6, 2024

Diabetic retinopathy (DR), a progressive eye disorder, can lead to irreversible vision impairment ranging from no DR severe DR, necessitating precise identification for early treatment. This study introduces an innovative deep learning (DL) approach, surpassing traditional methods in detecting stages. It evaluated two scenarios training DL models on balanced datasets. The first employed image enhancement via contrast limited adaptive histogram equalization (CLAHE) and generative adversarial network (GAN), while the second did not involve any enhancement. Tested Asia pacific tele-ophthalmology society 2019 blindness detection (APTOS-2019 BD) dataset, enhanced model (scenario 1) reached 98% accuracy 99% Cohen kappa score (CKS), with non-enhanced 2) achieving 95.4% 90.5% CKS. combination of CLAHE GAN, termed CLANG, significantly boosted model's performance generalizability. advancement is pivotal intervention, offering new pathway prevent loss diabetic patients.

Язык: Английский

---

Exploring open source and proprietary LoRa mesh technologies

Indonesian Journal of Electrical Engineering and Computer Science, Год журнала: 2024, Номер 34(2), С. 960 - 960

Опубликована: Март 23, 2024

This paper explores low power wide area network (LPWAN) LoRa and its diverse variants, encompassing open-source proprietary wireless mesh networks, operating over the physical or LoRaWAN layer. The primary challenge lies in defining an optimal solution that balances cost-effectiveness, energy efficiency, latency, long-range capability, security. study also comprehensively examines key solutions from 2017 to 2024, as proposed by various authors. Furthermore, a detailed analysis is conducted contrast commercial solutions, considering their applications, limitations, issues, characteristics, pros cons of routing protocols. In current landscape, proliferation has been instrumental facilitating connectivity internet things (IoT) devices. However, these pose challenges related consumption, suboptimal transmission throughput. These are influenced characteristics such spectrum factor, bandwidth, power, which directly impact range. Our research aims perform comparative existing by, systematically studying advantages disadvantages. offers valuable insights for making informed choices among domains IoT applications.

Язык: Английский

---