Indonesian Journal of Electrical Engineering and Computer Science
Not a member yet
9109 research outputs found
Sort by
Early skin disease diagnosis by using artificial neural network for internet of healthcare things
Get PDFInternet of healthcare things (IoHT) represents a burgeoning field that leverages pervasive technologies to create technology driven environments for healthcare professionals, thereby enhancing the delivery of efficient healthcare services. In remote and isolated areas, such as rural communities and boarding schools, access to healthcare professionals (especially dermatologists) can be particularly challenging. However, these areas often lack the specialized expertise required for effective skin disease consultations. Thus, the purpose of this research is to design a scheme of early skin disease diagnosis for internet of healthcare things that is accessible anywhere and anytime. In this research, the image of skin disease from patient will be taken by using a mobile phone for predicting and identifying the disease. This proposed scheme will diagnose skin disease and convert it be meaningful information. As a result, it show our proposed scheme can be the most consistent in term of accuracy and loss compared to others method. Overall, this research represents a significant step toward improving healthcare accessibility and empowering individuals to manage their own health. Furthermore, the proposed scheme is anticipated to contribute significantly to the IoHT field, benefiting both academia and societal health outcomes
Solar-powered irrigation and monitoring system for okra cultivation
Get PDFEco-friendly and cost-effective irrigation systems are essential for sustainable agriculture. Traditional irrigation systems are unsustainable due to the high cost of operation and environmental pollution associated with fossil fuels. A possible solution for farmers is the use of solar-powered irrigation systems. This research aims to develop a solar-powered irrigation and a real-time monitoring system for okra cultivation. The irrigation system was powered by a monocrystalline solar panel and controlled by a Node MicroController Unit ESP8266 microcontroller unit. A 12 V pneumatic diaphragm water pump was utilized to irrigate the okra plants efficiently. The real-time monitoring system using Blynk allowed for the remote monitoring of the system's performance. The irrigation system was deployed on an okra farm, and the results showed that the system could sustain the soil moisture level for the okra plants, with an average soil moisture sensor reading of over 80%. The system delivered power effectively, with an average voltage measurement exceeding 12 V, average current readings above 180 mA, and average power readings exceeding 2 W. These results demonstrate that the solar-powered irrigation system is a viable and sustainable solution for farmers, researchers, and engineers to enhance the performance of conventional irrigation systems
Tomato leaf disease detection using Taguchi-based Pareto optimized lightweight CNN
Get PDFThe prospect of food security becoming a global danger by 2050 due to the exponential growth of the world population. An increase in production is indispensable to satisfy the escalating demand for food. Considering the scarcity of arable land, safeguarding crops against disease is the best alternative to maximize agricultural output. The conventional method of visually detecting agricultural diseases by skilled farmers is time-consuming and vulnerable to inaccuracies. Technology-driven agriculture is an integral strategy for effectively addressing this matter. However, orthodox lightweight convolutional neural network (CNN) models for early crop disease detection require fine-tuning to enhance the precision and robustness of the models. Discovering the optimal combination of several hyperparameters might be an exhaustive process. Most researchers use trial and error to set hyperparameters in deep learning (DL) networks. This study introduces a new systematic approach for developing a less sensitive CNN for crop leaf disease detection by hyperparameter tuning in DL networks. Hyperparameter tuning using a Taguchi-based orthogonal array (OA) emphasizes the S/N ratio as a performance metric primarily dependent on the model’s accuracy. The multi-objective Pareto optimization technique accomplished the selection of a robust model. The experimental results demonstrated that the suggested approach achieved a high level of accuracy of 99.846% for tomato leaf disease detection. This approach can generate a set of optimal CNN models’ configurations to classify leaf disease with limited resources accurately
A hybrid combination of improved mayfly optimization based modified perturb and observe for solar based water pumping system
Get PDFIn recent years, solar water pumping systems (WPS) have been fuel-free and environmentally beneficial because they have gained a lot of attention in the agricultural and industrial sectors. Traditional water pumps consume higher amount of energy which make it as frequently unreliable, low efficiency and needs high maintenance. For WPS applications, Brushless DC (BLDC) motors are far superior options than other induction motors because of their high efficiency, high dependability, and low maintenance needs. Thus, in this research, the major goal is to develop a more efficient, reliable, and maintenance-free solar WPS solution. This paper describes a sensorless control strategy that reduces the need for hall sensors and increases system’s overall reliability. Solar system power is typically impacted by partial shadowing and cannot reach the maximum available power because the traditional perturbed and observe (P&O) algorithm fails. This paper integrates the modified P&O (MP&O) algorithm with an improved mayfly optimization (IMO) name called IMO-MP&O to address these issues by efficiently extracts the maximum power from solar. From the results, it clearly shows that IMO-MP&O achieved higher efficiency of 99.58% than the existing P&O MPPT which is analyzed the MATLAB sim-power-system toolboxes
Leveraging image fusion and transfer learning for enhanced tumor diagnosis
Get PDFIn India, the number of brain tumor cases are increasing rapidly. Compared to the current treatment approaches there is a need for efficient and advanced diagnostic approaches. Doctor primarily use magnetic resonance imaging (MRI) and computed tomography (CT) scans for diagnosis, each with its own set of advantages and limitations. Accurate diagnosis of brain tumors requires detailed information about the tumor’s type, size, location, and proliferation rate. But all this information must be estimated accurately and precisely. Even though MRI and CT provide anatomical information regarding tumors, they may not always correctly classify the tumor hence, often requiring the additional biopsy for more detailed analysis. This paper aims to demonstrate the fusion of MRI and CT scans by generating a fused image using transfer learning with convolution neural networks (CNN) that possesses all the important information from both modalities yields better results than using MRI and CT scans alone. This fused scan has the potential to highlight subtle details that may be overlooked on individual scans. This proposed system retains all the best possible functionalities for medical approaches to brain tumor detection. It paves the way to offer a much more impactful healthcare approach to patients by delivering accurate and reliable statistics, enhancing the diagnostic accuracy (96.43%) precedent to any single modality
A joint learning classification for intent detection and slot filling with domain-adapted embeddings
Get PDFFor dialogue systems to function effectively, accurate natural language understanding is vital, relying on precise intent recognition and slot filling to ensure smooth and meaningful interactions. Previous studies have primarily focused on addressing each subtask individually. However, it has been discovered that these subtasks are interconnected and achieving better results requires solving them together. One drawback of the joint learning model is its inability to apply learned patterns to unseen data, which stems from a lack of large, annotated data. Recent approaches have shown that using pretrained embeddings for effective text representation can help address the issue of generalization. However, pretrained embeddings are merely trained on corpus that typically consist of commonly discussed matters, which might not necessarily contain domain specific vocabularies for the task at hand. To address this issue, the paper presents a joint model for intent detection and slot filling, harnessing pretrained embeddings and domain specific embeddings using canonical correlation analysis to enhance the model performance. The proposed model consists of convolutional neural network along with bidirectional long short-term memory (BiLSTM) for efficient joint learning classification. The results of the experiment show that the proposed model performs better than the baseline models
Hybrid resource optimization strategy in heterogeneous wireless networks
Get PDFThe future generation of heterogeneous wireless networks (HWNs) will combine various radio access technologies for connecting various mobile subscribers (MS) based on the quality of service (QoS) and wireless network parameters, connecting MS to the best possible wireless network (WN) has been a trending research topic in HWNs. Existing resource optimization methods are designed to meet the QoS of network criteria and user preferences are neglected. Very limited work is done for resource optimization considering user preferences. However, these models are designed considering multi-mode terminals (MMTs) running a single service at a time under a low-density network; as a result, cannot be adopted to run multiple services simultaneously and; thus, fail to meet current users’ service dynamics requirement. Further, fails to bring good tradeoffs between reducing interference and improving performance. In addressing the research problem this work introduced a hybrid resource optimization strategy (HROS) to reduce interference by establishing channel availability and enhancing resource utilization through game theory. The HROS proves the existence of nash equilibrium (NE) improves throughput by 16.32% and reduces collision by 26.16% over the existing resource optimization-based network selection (RONS) scheme
Lung cancer prediction with advanced graph neural networks
Get PDFThis research aims to enhance lung cancer prediction using advanced machine learning techniques. The major finding is that integrating graph convolutional networks (GCNs) with graph attention networks (GATs) significantly improves predictive accuracy. The problem addressed is the need for early and accurate detection of lung cancer, leveraging a dataset from Kaggle's "Lung Cancer Prediction Dataset," which includes 309 instances and 16 attributes. The proposed A-GCN with GAT model is meticulously engineered with multiple layers and hidden units, optimized through hyperparameter adjustments, early stopping mechanisms, and Adam optimization techniques. Experimental results demonstrate the model's superior performance, achieving an accuracy of 0.9454, precision of 0.9213, recall of 0.9743, and an F1 score of 0.9482. These findings highlight the model's efficacy in capturing intricate patterns within patient data, facilitating early interventions and personalized treatment plans. This research underscores the potential of graph-based methodologies in medical research, particularly for lung cancer prediction, ultimately aiming to improve patient outcomes and survival rates through proactive healthcare interventions
Marine scientific workflow execution resource provisioner in edge-cloud platform
Get PDFWorkload application for deadline-intensive marine science, usually represented by directed acyclic graphs (DAGs), consists of interconnected operations that communicate huge quantities of information and operate on cutting-edge computational platforms. Massive communications of information among jobs running on separate computing servers, nonetheless, might come with substantial processing time, power consumption, and financial expenses. Therefore, there is room for further study of exchanging certain communications for computing to lower total interaction expenses. In addressing research this work introduces an effective resource provisioner for deadline-intensive scientific workload executer (ERP-DISWE) to reduce the makespan, consumption of energy, and cost of edge-computing platforms. The performance of ERP-DISWE is validated with the current resource provisioner using the sipht marine workload and shows superior performance concerning makespan, energy, and cost
Modeling non-linear communication systems using neural networks
Get PDFNonlinear systems present significant modeling challenges due to their complex dynamics and often unpredictable behavior. Traditional mathematical approaches can struggle to represent such systems accurately. In recent years, neural networks have emerged as promising tools to address this challenge. This article explores the use of neural networks to model nonlinear systems, focusing specifically on the application of the Hammerstein system. We examine network architecture and training methodologies suited to the complexity of nonlinear dynamics. Additionally, we explore strategies to improve the interpretability of neural network models in this context, enabling a better understanding of the underlying behavior of the system. Through a case study and empirical evaluations, we demonstrate the effectiveness of neural network-based approaches for estimating the behavior of nonlinear systems. Our work highlights the potential of neural networks as a versatile and powerful tool for modeling complex nonlinear phenomena