Indonesian Journal of Electrical Engineering and Computer Science
Not a member yet
9109 research outputs found
Sort by
Vehicle recognition on indian roads using data augmentation and VGG-16 model
Get PDFIn an advanced intelligent transportation system vehicle recognition and classi f ication is very significant. In current research trend, recognition of vehicles is done byusingmachinelearning (ML)andcomputervisiontechniques. Vehicle’s multi-view images or videos with different lighting conditions are annotated and given to the deep neural network to build an automated system to recognize the vehicles models. The augmentation of data can increase the number of sam ples in learning, with the small available datasets. Geometric transformations, brightness changes, and different filter operations are applied to the data through data augmentation. Furthermore, be orthogonal experiments we determine the optimal data augmentation method to obtain 96% accuracy in results. Detailed information is reported based on the classification of four different types of vehi cles and the results show that convolutional neural network with 16 layers deep techniques are effective in solving challenging tasks while recognizing moving vehicles
Artificial intelligence in diagnostic medicine: a case study of kidney disease applications
Get PDFThe rapid evolution of artificial intelligence (AI), particularly in convolutional neural networks (CNNs) and deep learning, has revolutionized numerous domains, ranging from medical imaging to creative arts and legal analytics. This research emphasizes the role of pre-trained CNN architectures in identifying kidney conditions, leveraging a dataset comprising images of healthy kidneys as well as those affected by cysts, tumors, and stones. The pretrained models known for their outstanding image recognition capabilities, were adapted for this classification task through transfer learning (TL) techniques. By refining these models and carefully calibrating key parameters like learning rate, batch size, and network depth, they demonstrated superior performance compared to traditional machine learning approaches. The findings underscore the transformative potential of pre-trained CNNs in advancing the precision of kidney disease diagnostics, with implications for broader medical applications
Evolving strategies in anti-phishing: an in-depth analysis of detection techniques and future research directions
Get PDFPhishing attacks are a major digital threat, impacting individuals and organizations globally. This review paper examines evolving anti-phishing strategies by analyzing five key techniques: URL blacklists, visual similarity detection, heuristic methods, machine learning models, and deep learning techniques. Each technique is evaluated for its mechanisms, unique features, and challenges. A systematic literature survey (SLR) is conducted to compare these methods; effectiveness. The paper highlights significant research challenges and suggests future directions, emphasizing the integration of artificial intelligence and behavioral analytics to combat evolving phishing tactics, this study aims to advance understanding and inspire more effective anti-phishing solutions
Exploration of various approaches for detection of autism spectrum disorder
Get PDFAutism spectrum disorder (ASD) presents a complex and diverse set of challenges, necessitating innovative and data-driven approaches for effective understanding, diagnosis, and intervention. This review explores recent advancements in methodologies, technologies, and frameworks aimed at addressing ASD and also highlights novel data collection methods, focusing on the integration of wearable internet of things (IoT) sensors for real-time behavioral monitoring and data capture from individuals with ASD. Additionally, the utilization of machine learning (ML), deep learning (DL), and hybrid techniques for data analysis, feature optimization, and prediction of ASD are extensively discussed, showcasing significant progress in early diagnosis and personalized intervention planning. The challenges such as class imbalance, feature selection, and data collection efficiency are identified and addressed using the proposed ASD framework. The review also emphasizes the development of recommendation systems designed to the unique behavioral profiles and needs of individuals with ASD. The findings reveal that integrating these advanced technologies and methodologies can lead to more accurate diagnoses and effective interventions, contributing to the broader field of ASD research
Compressor performance prediction: gradient boosting regression model and sensitivity analysis
Get PDFThis study introduces the use of gradient boosting regression (GBR) models to estimate the compressor performance of aero-engines. The model exhibits a mean absolute error (MAE) of 0.078, showcasing superior performance compared to previous studies. Through sensitivity analysis, optimal values for three key parameters were determined: 280 estimators, a max depth of 9, and a learning rate of 0.085. Furthermore, a comparison with a prior study revealed an impressive MAE value lower than 0.002, highlighting the GBR model’s success in accurately predicting compressor performance. This demonstrates the model’s effectiveness and predictive accuracy, making it a valuable tool for aero-engine compressor performance estimation
IT risks associated with information theft in the financial system - a systematic review
Get PDFThis research paper systematically reviews the financial system’s computer security risks associated with information theft. The objective is to explore the security risks and their implications concerning information theft in the economic system. Three research questions were formulated to identify these risks, their nature, and potential consequences to achieve this objective. Fifty-five articles obtained from reliable databases linked to both study variables were analyzed using the PRISMA methodology. To ensure the validity and reliability of the information, various filters were applied, such as year, keywords, and elimination of duplicate articles. In addition, an exhaustive reading of the content of each article was carried out, organizing all the information through a systematization matrix. After a thorough review of the research articles, mostly written in English and representing 34.55% of the total in 2023, risks associated with the financial sector were identified, including malware, ransomware, phishing, distributed denial of service (DDoS), hybrid XSS, eavesdropping, and social engineering. Geographically, India leads with 14.55% of the articles, followed by South Korea and the United States, with 12.72% each, while the other countries have lower percentages. In conclusion, these risks coincide with previous research and the consequences they generate, highlighting the importance of this type of study for the basis of scientific research
A comprehensive overview of LLM-based approaches for machine translation
Get PDFStatistical machine translation (SMT) used parallel corpora and statistical models, to identify translation patterns and probabilities. Although this method had advantages, it had trouble with idiomatic expressions, context-specific subtleties, and intricate linguistic structures. The subsequent introduction of deep neural networks such as recurrent neural networks (RNNs), long short-term memory (LSTMs), transformers with attention mechanisms, and the emergence of large language model (LLM) frameworks has marked a paradigm shift in machine translation in recent years and has entirely replaced the traditional statistical approaches. The LLMs are able to capture complex language patterns, semantics, and context because they have been trained on enormous volumes of text data. Our study summarizes the most significant contributions in the literature related to LLM prompting, fine-tuning, retrieval augmented generation, improved transformer variants for faster translation, multilingual LLMs, and quality estimation with LLMs. This new research direction guides the development of more efficient and innovative solutions to address the current challenges of LLMs, including hallucinations, translation bias, information leakage, and inaccuracy due to language inconsistencies
Quantitation of new arbitrary view dynamic human action recognition framework
Get PDFDynamic action recognition has attracted many researchers due to its applications. Nevertheless, it is still a challenging problem because the diversity of camera setups in the training phases are not similar to the testing phases, and/or the arbitrary view actions are captured from multiple viewpoints of cameras. In fact, some recent dynamic gesture approaches focus on multiview action recognition, but they are not resolved in novel viewpoints. In this research, we propose a novel end-to-end framework for dynamic gesture recognition from an unknown viewpoint. It consists of three main components: (i) a synthetic video generation with generative adversarial network (GAN)-based architecture named ArVi-MoCoGAN model; (i) a feature extractor part which is evaluated and compared by various 3D CNN backbones; and (iii) a channel and spatial attention module. The ArVi-MoCoGAN generates the synthetic videos at multiple fixed viewpoints from a real dynamic gesture at an arbitrary viewpoint. These synthetic videos will be extracted in the next component by various three-dimensional (3D) convolutional neural network (CNN) models. These feature vectors are then processed in the final part to focus on the attention features of dynamic actions. Our proposed framework is compared to the SOTA approaches in accuracy that is extensively discussed and evaluated on four standard dynamic action datasets. The experimental results of our proposed method are higher than the recent solutions, from 0.01% to 9.59% for arbitrary view action recognition
Correlation between input and output parameters of microbial fuel cell
Get PDFThis paper presents the correlation between open circuit voltage (OCV) and pH, temperature, and total dissolved solids (TDS) of an air cathode single chamber microbial fuel cell (MFC) using artificial neural network (ANN) and support vector machine (SVM) algorithms. Previous works used terminal voltages as output parameters to determine the correlation between MFCs' input and output parameters. However, OCV is the most important measurement that can determine the validity of the MFC. Thus, various tests were conducted to analyze the correlation between OCV and input parameters using ANN and SVM algorithms. Both techniques show a strong correlation between OCV and input parameters with the highest R2 values. The highest OCV value obtained from the experiment is 1.179 V at pH 5.26, temperature 299K, and TDS 3,124 ppm. Furthermore, an ANN model was developed to predict the OCV value based on pH, temperature, and TDS value
Empower BreastNet: breast cancer detection with transfer learning VGG Net-19
Get PDFBreast cancer is a major cause of death among women globally, making early detection crucial for effective treatment. This study introduces a new deep learning (DL) method using transfer learning (TL) to automatically detect and diagnose breast cancer. TL improves performance on new tasks by using knowledge from previous tasks. In this study, we use pre-trained convolutional neural networks (CNNs) like AlexNet, ResNet50, visual geometry group (VGG)-16, and VGG-19 to extract features from the breast cancer wisconsin (BCW) diagnostic dataset. We measure the model's success with accuracy, sensitivity, specificity, precision, and F-score. The results show that the VGG-19 model, when applied with TL, performs best for diagnosing breast cancer, achieving an overall accuracy of 98.75%, sensitivity of 97.38%, specificity of 98.35%, precision of 97.35%, and an F-score of 97.66%