Indonesian Journal of Electrical Engineering and Computer Science
Not a member yet
9109 research outputs found
Sort by
Combination of MLF-VO-F and loss functions for VOE from RGB image sequence using deep learning
Get PDFVisual odometry estimation (VOE) is important in building navigation and pathfinding systems. It helps entities find their way and estimate paths in the environment. Most of the computer vision (CV)-based VOE models are usually evaluated and compared on the KITTI dataset. Multi-layer fusion framework (MLF-VO-F) has had good VOE results from red, green, and blue (RGB) image sequence in Jiang et al. study, using the DeepNet to extract the low-level textures, edges, and deeper high-level semantic features for estimating motion between consecutive frames. This paper proposed a combined model of MLFVO-F as a backbone and loss functions (LFs) (LMSE, LMSE−L2, LCE, and Lcombi) to optimize and supervise the training process of the VOE model. We evaluated and compared the effectiveness of LFs for VOE based on the KITTI and TQU-SLAM datasets with the original MLF-VO-F. From there, choose the appropriate LF combined with the backbone for VOE. The evaluation results on the KITTI dataset show that LCE(RT E is 0.075m, 0.06m on the Seq. #9, Seq. #10, respectively), and Lcombi (trel is 2.21%, 2.67%, 3.59%, 1.01%, and 4.62% on the Seq. #4, Seq. #5, Seq. #6, Seq. #7, Seq. #10, respectively) have the lowest errors and LMSE has the highest errors (AT E is 133.36m on the Seq. #9)
An efficient load balance using virtual machine migration hybrid optimization technique in cloud computing
Get PDFCloud computing is becoming increasingly important to developers and companies because to the rapid development of information technology and the wide availability of internet applications. Every information technology industry has a significant role for cloud computing. Numerous multinational technology businesses, like Google, Microsoft, and Facebook, have established data centers across the world to offer processing and storage capabilities. Customers can submit their jobs to cloud centers directly. Reducing overall power usage is the primary goal, which was overlooked in the early stages of cloud development. Using gene expression programming (GEP), symbolic regression models of virtual machines (VMs) are developed using measured VM loads and the corresponding resource parameters. In order to minimize resource use, multidimensional resource load balancing of all the physical machines within the cloud computing platform is the aim of this analysis. The VMH loads estimated and the genetic algorithm that considers the current and the future loads of VMHs and decides an optimal VM-VMH for migrating VMs and performing load-balance. Hence, an efficient load balance using virtual machine migration hybrid optimization technique (HOT) in cloud computing shows better results in terms of accuracy, energy consumption, migration cost
Brain tumor classification for optimizing performance using hybrid RNN classifier
Get PDFTumor is the uncontrolled growth of cancer cells in any part of the human body. Brain tumoris the leading cause of cancer deaths worldwide among adults and childrens. Early detection of brain cancers is essential. To prevent more issues, early defect detection is essential. Healthcare physicians may discover and categorize brain tumors with the use of computational intelligence-focused tools. An essential task for diagnosing tumors and choosing the right type of therapy is classifying brain tumors. Brain tumor identification and segmentation using magnetic resonance imaging (MRI) scans is now recognized as one of the most significant and difficult research areas in the world of medical image processing. The field of medical imaging has gained greatly from the use of artificial intelligence (AI) in the form of machine learning (ML) and deep learning (DL). DL has shown significant presentation, especially in the areas of brain tumor classification and segmentation. In this work, brain tumor classification for optimizing performance using hybrid recurrent neural network (RNN) classifier is presented. Different types of brain tumors are classified using a mix of RNN and inception residual neural network (ResNet). This strategy will produce improved F1-score, precision, accuracy, and recall scores
Deep learning-based multi-tier sensitivity analysis network for document sensitivity classification
Get PDFIn the digital age, the exponential growth of data necessitates robust and efficient systems for document classification to maintain data security and compliance. Text classification plays a crucial role in identifying sensitive information by automatically categorizing documents based on their content. Using advanced machine learning and deep learning models, it analyzes text to detect keywords, patterns, and contextual cues that indicate the presence of sensitive data. This paper presents a novel framework, the multi-tier sensitivity analysis network (MTSAN), designed to accurately classify documents into public, private, and confidential categories. The proposed system integrates several advanced components, including the multi-tier sensitivity encoding network (MTSEN). MTSAN leverages a combination of convolutional networks and graph convolutional networks (GCNs) to capture both local and global contextual information. The dual-scope graph convolution block (DSGCB) is introduced to address both global dependencies and local dynamics, employing a novel fusion mechanism to merge global and local features effectively. Additionally, the cross-tier information fusion block (CTIFB) facilitates the seamless integration of multi-level features, further refining the classification process. The results demonstrate that the proposed MTSAN model outperforms traditional machine learning approaches and contemporary deep learning models such as bidirectional encoder representations from transformers (BERT), achieving superior accuracy and F1 scores in classifying sensitive information
Digital and academic libraries through cloud computing
Get PDFIn an era characterized by the dominance of digital information, libraries have undergone significant transformations, evolving from traditional brickand-mortar institutions to dynamic hubs of digital knowledge. The emergence of digital libraries, which give users access to vast collections of digital resources, has facilitated this evolution. However, effective management of digital resources poses numerous challenges, including issues related to storage, preservation, and accessibility. In response, cloud computing has developed as a powerful solution for addressing these challenges and revolutionizing how libraries operate. Cloud computing reduces the need for expensive infrastructure expenditures and increases flexibility and scalability by allowing libraries to store, manage, and access digital resources remotely over the internet. This paper examines the intersection of digital libraries and cloud computing, examining the role of cloud computing in modern libraries and its implications for the future of information management. By analyzing current trends, case studies, and best practices, this paper provides insights into the benefits and challenges of adopting cloud computing in the context of academic libraries
Advanced deep attention neural inference network for enhanced arrhythmia detection and accurate classification
Get PDFArrhythmias are irregular heartbeats that can lead to severe health risks, including sudden cardiac death, necessitating accurate and timely detection for effective treatment. Traditional diagnostic methods such as stress tests, resting electrocardiograms (ECGs), and 24-hour Holter monitors are limited by their monitoring capacity and often result in delayed diagnoses, compromising patient safety. To address these challenges, this paper introduces the deep attention neural inference network (DANIN) methodology. DANIN integrates one-dimensional ECG signals with two-dimensional spectral images using multi-modal feature fusion, capturing comprehensive cardiac information in both temporal and frequency domains. The methodology employs advanced deep attention network-based models for superior feature extraction, recognizing intricate patterns and long-range dependencies within the data. Additionally, the inclusion of an inference model system enhances interpretability and usability, making the model highly suitable. Further, DANIN is evaluated considering the MIT-BIH dataset, and extensive comparative analysis with state-of-the-art techniques demonstrates that DANIN significantly improves accuracy, precision, recall, and F1-score, highlighting its potential to revolutionize arrhythmia detection and improve patient outcomes
Effective methods for employee performance assessment
Get PDFThis study aims to select the most effective multi-criteria decision-making method used in an employee performance appraisal system. The approach used in this study is a comparative experiment where three multi-criteria decision-making methods simple additive weighting (SAW), analytical hierarchy process (AHP), and technique for order preference similarity to an ideal solution (TOPSIS) are compared. The dataset involves 16 employees, considering input data such as work behavior scores, and performance targets (SKP). The criteria for evaluating work behavior include service quality, accountability, competence, harmony, loyalty, adaptability, collaboration, and achievement of targets. The comparison results were tested using a one-way ANOVA to evaluate whether there are significant differences among the three methods, as well as to provide supporting evidence for the conducted research. The results indicated that the SAW method provides the most accurate and relevant performance assessments while AHP yields less precise rankings as some employees received the same scores despite having different workloads. TOPSIS also produced rankings that did not accurately reflect the relative workloads. Implementing the SAW method in the employee performance information system enhances the assessment process, making it faster, more objective, transparent, and credible. Thus, SAW emerges as the most effective method for aligning performance scores with employee roles and responsibilities
Autonomous driving system and system hacking protection using V2X communication
Get PDFIn this paper, a new autonomous driving system is proposed and problems such as systematic errors that may occur in the autonomous driving system were solved through vehicle to everything (V2X) communication technology. In the actual driving environment, accidents caused by the absence of communication between drivers and communication with infrastructure are frequently exposed. To solve these problems, a system was established that linked V2X communication with a vehicle system. In order to predict and study how this technology works in real traffic situations, it requires a lot of time, manpower, and funds because it requires building an environment similar to real traffic situations and using measuring equipment. For this reason, the system was built with simple model, and the research was conducted through simple simulation. In addition, as network technology and sensing technology for autonomous vehicles develop, the risk of hacking is also increasing. In this paper, various expected attack paths and methods that can hack autonomous vehicles are explained, and methods for defending them are presented
Gender identification from tribal speech using several learning techniques
Get PDFLanguage processing and linguistics researchers are interested in gender identification through audio, as human voices have many distinctive features. Although several gender identification algorithms have been developed, the accuracy and efficiency of the system can still be improved. Despite extensive studies on the topic in various languages, there aren’t many studies on gender identification in the KUI language. Using a variety of machine learning (ML) and deep learning (DL) classifiers, including decision tree (DT), multilayer perceptron (MLP), gradient boosting (GB), linear discriminant analysis (LDA), recurrent neural networks (RNN), long short-term memory (LSTM), gated recurrent units (GRU), and transformer, the goal of this study is to assess the accuracy of gender identification among diverse KUI language speakers. To verify the effectiveness of the suggested model, several prediction evaluation metrics were calculated, such as the area under the receiver operating characteristic curve (AUC), F1-score, precision, accuracy, and recall. While the findings are compared to other learning models, the gradient-boosting strategy yielded better results with an accuracy rate of 97.0%
Intelligent transportation network-based congestion forecasting with federated learning and a convolutional neural network
Get PDFThe heavy traffic in growing cities hurts the environment, commuters, and economy. Predicting such difficulties early helps increase road network capacity and efficiency and reduce congestion. Many academicians and transportation engineers ignore traffic congestion prediction despite its importance. Insufficient computationally efficient traffic forecast systems and high-quality city-wide traffic data contribute to this. Provide useful information to reduce traffic and construct shorter, more energy-efficient routes. Data storage increases traditional traffic forecasting training, storage costs, and delay. Smarter algorithms can handle today’s city expectations because sensors can now communicate with their environment. A vibrant economy requires decent roads. Improving transportation requires uninterrupted highway traffic. To overcome these issues, smart city roadway traffic flow must be monitored in real time using enhanced internet of things (IoT) capabilities. Training data may contain sensitive information, raising privacy problems. This work addresses these issues by training the prediction model near data sources using federated learning (FL). The suggested strategy was tested using Mumbai, Chennai, and Bangalore traffic data. We compared the proposed method to centralized strategies to assess its efficacy. Our experiments confirm the model’s traffic jam prediction accuracy. Our approach outperforms auto-encoder and convolutional neural network (CNN) in computer efficiency and prediction