Indonesian Journal of Electrical Engineering and Informatics (IJEEI)
Not a member yet
776 research outputs found
Sort by
Remote Sensing for Forensic Investigations: A Review of Techniques and Applications in Clandestine Grave Detection
Full text linkThe location of clandestine graves is a critical challenge in forensic investigations. This review evaluates the appli-cation of remote sensing technologies to address this challenge. A comprehensive literature search was conducted across scientific databases (Web of Science, Scopus, IEEE Xplore, Google Scholar, PubMed Central, ScienceDirect), using keywords related to remote sensing, forensic science, and burial detection. Peer-reviewed articles and books focusing on remote sensing applications in forensic contexts, especially clandestine grave detection, were included. Data on methods, location, target, spectral indices, and key findings were extracted. A significant increase in pub-lications in this field was observed, particularly since 2018. Techniques included multispectral and hyperspectral imaging (satellite and UAV), LiDAR, GPR, ERT, and thermal imaging. Spectral indices (NDVI, GNDVI, VARI) were used to analyze vegetation stress. Success varied with burial depth, soil type, vegetation cover, and time since burial. Geophysical methods provided valuable subsurface information, but effectiveness decreased over time. Remote sensing offers powerful tools for forensic investigations, enabling non-invasive assessment and improved detection of clandestine graves. A multidisciplinary approach, combining multiple remote sensing techniques with geophysical methods, is crucial. Further research is needed to optimize techniques for diverse environments, improve detection of older burials, and develop standardized methodologie
Mitigating Wormhole Attacks’ Risks within Wearable Body Network
Full text linkIn this research, we sought to develop a trust and secure routing protocol based on the Ad-hoc On-Demand Distance Vector (AODV) routing to combat wormhole attacks in Wearable Body Networks (WBNs), which integrates a routing strategy that leverages the path-checking method to detect and isolate paths affected by wormhole attacks effectively, it employs a routing technique that prioritizes nodes with the most heightened remaining energy during data transmission, along with a mixed cryptographic algorithm that combines the modified One Pad Time with the modified Affine ciphers to ensure safe transmission against malicious biosensor threats. Experimental findings indicate that our proposed protocol transcends the classic AODV routing protocol across all evaluation parameters, including packet delivery ratio, throughput, and energy consumption. Its primary advantage lies in considering multiple factors, like detecting unauthorized biomedical biosensors, efficient energy utilization in the network, and secure data transmission—differentiating it from other safe routing protocols. Moreover, the mixed encryption algorithm enhances efficacy and bolsters sensitive data security compared to classic cipher methods like the One Pad Time and Affine ciphers
Benchmarking of OFDM Spectrum Exchange for Mobile Cognitive Radio Networks
Full text linkThe local spectrum sensing objective in spectrum sensing is to detect the PU's signal. The sensing node's (SN) capacity to detect the PU's signal is of paramount importance. However, it is presumed to be stationary in the majority of SN in cognitive radio networks. The detection performance on local observation is significantly influenced by the mobility of the PUs and SNs. The SNs' movement generates spatial diversity in the PU's signal observation. The signal's condition would fluctuate during the sensing process as a result of Doppler effect, spatial distance, velocity, movement, and geolocation information. Therefore, a benchmark is required to compare the primary user signal detection level of stationary and moving SNs from each sensing node. The performance results have demonstrated that static nodes with SCM are superior to conventional subcarrier mapping (SCM) methods in the case of a subcarrier mapping width of α = 2. Additionally, the quantization width is uniform. It has been determined that the performance disparity is substantial, ranging from 2 dB to 4 dB. The results indicate that the static nodes SCM have achieved acceptable performance detection at a low subcarrier detection threshold (SDT) value of 0 dB up to 5 dB. Conversely, the probability of conventional SCM detection is less than 1 of probability detection (PD) value at the same low SDT value. The detection probability (PD) of static nodes with SCM is satisfactory at an SDT value of 15 dB. Moreover, the probability begins to decline until 20 dB at an SDT value of 11.5 dB, a substantial decrease that is rendered negligible. In contrast to the new subcarrier mapping (N-SCM) method, which has a false alarm probability (PFA) of approximately 0 dB to 9.5 dB, conventional subcarrier mapping (SCM) has a high false alarm probability in mobility networks. Furthermore, it is evident that the PFA curves for the conventional SCM method are lower than those of other methods at low speeds, as they approach the null value at SDT 7.5 dB. The PFA curve for both methods is higher than other velocities by attaining a null value at 10 dB, in contrast to high velocity. In general, the mobility parameter has the potential to meet the detection performance and perform well in the false alarm probability of mobile spectrum exchange. Consequently, it could be employed to provide information on spectrum exchange in the future
Using The Combined Model Between MobileNetV2 and EfficientNetB0 to Classify Brain Tumors Based on MRI Images
Full text linkBrain tumors are extremely dangerous to one's health. If unchecked cell proliferation is not identified and treated promptly, it can lead to mortality, raise intracranial pressure, and endanger lifespan. To remove the tumor and lengthen the patient's life, early illness identification and drug administration are essential. In this research paper, we aim to improve the effectiveness of magnetic resonance imaging (MRI) equipment to identify cancerous brain tumour cells. It helps experts identify diseases faster. We classify brain tumour cells based on an image set of 3264 images with effective classification models such as ResNet50, InceptionV3, VGG19, EfficientNetB7, DenseNet201, MobileNetV2, Xception, etc. Besides, we also proposed two combined models: pooling (Xception + ResNet50) and pooling (MobileNetV2 + EfficientNetB0) to evaluate the effectiveness and found that the pooling model (MobileNetV2 + EfficientNetB0) gives the highest result, with 100% for the training set, 98% for the valid set, and 78% for the test set. We continued to improve the model by randomly re-dividing the data set with a Train-Valid-Test ratio of 60:20:20 and obtained an increased F1-score of 97%. We continued to improve the model again using the data augmentation techniques to create a larger data set, and the results far exceeded expectations with an F1-score of almost 100% for all classes. Based on the results, we found that combining MobileNetV2 with EfficientNetB0 is suitable for detecting brain tumour cancer cells. Aids in the early detection of dangerous cancers before they spread and endanger human health
Development of a Network Intrusion Detection Model using Hybridised Machine Learning Algorithms
Full text linkCyber threats continue to grow in this era since the bad actors are attempting to exploit individuals, organisations, and systems. The latest development in artificial intelligence has unleashed strong agents at the fingertips of humanity. As open as it is, it has made more room for possible bad actors. Systems that can successfully counter these threat actors need to be created to rescue humanity. In this research work, RNN and Random Forest classifiers' hybridised models are combined for the development of a Network Intrusion Detection System (NIDS) based on the benchmark dataset (CICIDS 2017) The requirement for an efficient and accurate method to detect network intrusions, both known and zero-day anomalies, is the primary problem considered. This research aims to enhance the accuracy and reliability of intrusion detection systems through a hybrid modelling approach. For evaluating the performance of the proposed model, various measures like accuracy, precision, recall, F1 measure, true positive rate, and true negative rate were employed. The hybrid model showed very good results with testing accuracy of 96.08%, precision of 96.0%, and recall of 96.0%, along with an F1 measure of 96.0%. The result of the experiment indicates that the model is effective and, when implemented, can detect and classify cyberattacks in modern environments
Ensemble Based Machine Learning Approach for Heart Disease Prediction
Full text linkEnsemble machine learning has developed into a strong approach for enhancing the precision and resilience of predictive models through the integration of various learning algorithms. This research presents an innovative ensemble classification framework employing a soft voting approach that combines three gradient boosting techniques XGBoost, LightGBM, and CatBoost to improve heart disease prediction efficacy. The model undergoes evaluation using four distinct datasets (Heart Attack Risk Prediction Dataset, Heart Attack Dataset, Cleveland Heart Disease dataset and Heart Disease Dataset) obtained from Kaggle and other repositories, each reflecting various populations and diagnostic variables. By implementing thorough preprocessing, careful feature selection, and even training-testing-validating splits, the system attains reliable and exceptional classification performance. Experimental findings reveal that the suggested ensemble approach greatly surpasses classic and standalone models, attaining flawless or nearly flawless accuracy on all datasets, reaching a peak accuracy of 100% on the first dataset, 98% on the second dataset, 100% on the third dataset and 98.4% on the fourth dataset. The framework's achievement underscores its viability for real world use in clinical decision support systems and emphasizes the efficiency of ensemble methods in medical diagnosis
A Systematic Literature Review (SLR) of Mirai Botnet Compromise Detection in Internet of Things (IoT) Network
Full text linkSince its invention, Mirai botnet has remained a significant concern in IoT network security. The botnet and its evolving variants are a major threat to professionals responsible for securing IoT infrastructures. The danger of the botnet is attributed to the fact that it has been utilized for the execution of numerous Distributed Denial of Service (DDoS) attacks on different network infrastructures in the past. Several researchers have proposed techniques in mitigating the effect of this botnet. This research systematically reviews existing detection techniques and evaluates how effective they are in mitigating Mirai botnet attacks between 2017 and 2024. Using PRISMA methodology, 177 articles were initially identified from Scopus, Springer Link, IEEE Xplore, and Web of Science in order to broaden the scope of the search. 27 studies passed the inclusion criteria, and were analyzed thereafter. Findings reveal a predominant reliance on AI-driven detection methods, such as LSTM and ensemble models, which demonstrate higher accuracy and scalability when compared to traditional techniques. This review also compares threat intelligence platforms like AlienVault, CrowdStrike, and Recorded Future, to assess their contributions to dynamic detection frameworks. Finally, the study explores research gaps and proposes future directions for developing scalable real-time detection systems integrating multi-source threat feed
Efficient Corrosion Detection on Metal Surface Using Deep Learning Technique
Full text linkThis study examines how deep learning models can improve corrosion detection, comparing YOLOv7 with its more advanced version, YOLOv8. Both models were trained on a diverse set of images showing different types and levels of corrosion on metal surfaces. Their performance was assessed using standard industry metrics, including accuracy, F1-score, recall, and mean average precision (mAP). The results clearly show that YOLOv8 outperforms YOLOv7 in all areas. It achieves higher recall, precision, and F1-score, demonstrating its improved ability to detect and classify corroded areas. Notably, YOLOv8 is better at identifying small or early-stage corrosion, which is crucial for timely maintenance. Additionally, it processes images faster than YOLOv7, making it more suitable for real-time applications. This study also suggests integrating YOLOv8 with robotic arms equipped for laser cleaning, allowing for automated and precise corrosion removal. This system could improve maintenance efficiency, reduce costs, and enhance the safety and reliability of infrastructure
EfficientNet Model for Multiclass Classification of The Correctness of Wearing Face Mask
Full text linkA face mask is essential for protecting individuals from the entry of infectious or hazardous materials through the nose or mouth in specific situations. To optimize its protective function, it must be worn correctly. This research aims to develop a multiclass classification model, rather than a binary one, to assess the correctness of wearing face mask. The proposed model is designed to achieve high accuracy while maintaining efficiency, with a low number of model parameters. To this end, a deep convolutional neural network (CNN), specifically EfficientNet, is utilized. Experiments are conducted on the public MaskedFace-Net image dataset, which consists of four categories (correctly masked, uncovered chin, uncovered nose, and uncovered nose and mouth), using 3,000 randomly selected images from each category. The experiments test several EfficientNet models (B0-B3) and network hyperparameters (learning rate and dropout). The best accuracy of 0.99 is achieved by EfficientNet-B0 with a learning rate of 0.01 and a dropout rate of 0.2. The EfficientNet-B0 model outperforms other benchmark CNN models, including MobileNet-V3 and Inception-V3, despite having a slightly higher number of parameters than MobileNet-V3. This result demonstrates that the EfficientNet model is both accurate and efficient for multiclass classification of the correctness of wearing face mask
Enhancing LEACH Protocol with Multi-Criteria Decision Making for Prolonged Network Lifetime in WSNs
Full text linkWireless Sensor Networks (WSNs) have become a crucial solution for monitoring across diverse environments and consist of tiny sensor nodes that autonomously gather data on the environment. Energy depletion is a looming challenge, as sensor nodes rely heavily on their batteries, and once exhausted, the entire network can collapse prematurely. The Low-Energy Adaptive Clustering Hierarchy (LEACH) protocol is a cornerstone in energy-efficient routing protocols for WSNs. However, the Cluster Head (CH) selection process in the traditional LEACH protocol relies on a probabilistic model for CH selection, where each sensor has an equal chance of becoming a CH based on a fixed threshold. To address these issues, this paper proposes an enhanced version of the LEACH protocol by employing a Multi-Criteria Decision-Making (LEACHMCDM) process for CH selection. Instead of relying on random probabilities, the proposed protocol incorporates three key factors: Residual Energy (RE), Distance to the Base Station (DBS), and Node Degree (ND). Nodes with higher RE, shorter DBS, and an optimal ND are more likely to be selected as CHs. Compared to the traditional LEACH, the proposed method significantly improves the network’s lifetime by evenly distributing energy consumption and reducing the risk of premature node failure. Simulation results demonstrate the enhanced protocol’s ability to sustain more operational rounds and achieve higher energy efficiency