JOIV : International Journal on Informatics Visualization
Not a member yet
786 research outputs found
Sort by
Batik Images Retrieval Using Pre-trained model and K-Nearest Neighbor
Get PDFBatik is an Indonesian cultural heritage that should be preserved. Over time, many batik motifs have sprung up, which can lead to mutual claims between craftsmen. Therefore, it is necessary to create a system to measure the similarity of a batik motif. This research is focused on making Content-Based Image Retrieval (CBIR) on batik images. The dataset used in this research is big data Batik images. The authors used transfer learning on several pre-trained models and used Convolutional Neural Network (CNN) Autoencoder from previous studies to extract features on all images in the database. The extracted features calculate the Euclidean distance between the query and all images in the database to retrieve images. The image closest to the query will be retrieved according to the number of r, namely 3, 5, 10, or 15. Before the image is retrieved, the retrieval system is used to re-ranked with K-Nearest Neighbor (KNN), which classifies the retrieved image. The results of this study prove that MobileNetV2 + KNN is the best model in terms of Image Retrieval Batik, followed by InceptionV3 and VGG19 as the second and third ranks. Moreover, CNN Autoencoder from previous research and InceptionResNetV2 are ranked fourth and fifth. In this study, it was also found that the use of KNN re-ranking can increase the precision value by 0.00272. For further research, deploying these models, especially for MobileNetV2 is an approach for seeing a major impact on batik craftsmanship for decreasing batik motif plagiarism
Roboswab: A Covid-19 Thermal Imaging Detector Based on Oral and Facial Temperatures
Get PDFThe SARS-CoV-2 virus has been the precursor of the coronavirus disease (COVID-19). The symptoms of COVID-19 begin with the common cold and then become very severe, such as those of Middle East Respiratory Syndrome (MERS) and Severe Acute Respiratory Syndrome (SARS). Currently, polymerase chain reaction (PCR) is used to detect COVID-19 accurately, but it causes some side effects to the patient when the test is performed. Therefore, the proposed "Roboswab" was developed that uses thermal imaging to measure non-contact facial and oral temperature. This study focuses on the performance of the proposed equipment in measuring facial and oral temperature from various distances. Face detection also involves checking whether the subject is wearing a mask or not. Image processing methods with thermal imaging and robotic manipulators are integrated into a contact-free detector that is inexpensive, accurate, and painless. This research has successfully detected masked or non-masked faces and accurately detected facial temperature. The results showed that the accurate measurement of facial temperature with a mask is 90% with an error of +/- 0.05%, while it was 100% without a mask. On the other hand, the oral temperature was measured with 97% accuracy and an error of less than 5%. The optimal distance of the Roboswab to the face for measuring temperature is an average of 60 cm. The Roboswab tool equipped with masked or non-masked face detection can be used for early detection of COVID-19 without direct contact with patients
Improving Badminton Player Detection Using YOLOv3 with Different Training Heuristic
No full textThere has been a considerable rise in the amount of research and development focused on computer vision over the previous two decades. One of the most critical processes in computer vision is "visual tracking," which involves following objects with a camera. Tracking objects is the practice of following an individual moving object or group of moving things over time. Identifying or connecting target elements in consecutive video frames of a badminton match requires visual object tracking. The aim of this study is to identify badminton players using the You Only Look Once (YOLO) technique in conjunction with a variety of training heuristics. This methodology has a few advantages over other approaches to detecting objects. The convolutional neural network and Fast convolutional neural network are two examples of the many algorithmic approaches that are available. In this study, a neural network is used to produce predictions about the bounding boxes and the class probabilities for these boxes.. The results demonstrated that it was far faster than other methods in terms of its ability to recognize the image. The performance of image classification networks significantly improved as a result of the implementation of a variety of training strategies for the detection of objects. The mean average precision score for YOLOv3 with various training heuristics increased from 32.0 to 36.0 as a direct result of these adjustments. In comparison to YOLOv3, our future study might examine the performance of alternative models like Faster R-CNN or RetinaNet
Vehicles Speed Estimation Model from Video Streams for Automatic Traffic Flow Analysis Systems
Get PDFImage and video processing have been widely used to provide traffic parameters, which will be used to improve certain areas of traffic operations. This research aims to develop a model for estimating vehicle speed from video streams to support traffic flow analysis (TFA) systems. Subsequently, this paper proposes a vehicle speed estimation model with three main stages of achieving speed estimation: (1) pre-processing, (2) segmentation, and (3) speed detection. The model uses a bilateral filter in the pre-processing strategy to provide free-shadow image quality and sharpen the image. Gaussian filter and active contour are used to detect and track objects of interest in the image. The Pinhole model is used to assess the real distance of the item within the image sequence for speed estimation. Kalman filter and optical flow are used to flatten vehicle speed and acceleration uncertainties. This model is evaluated with a dataset that consists of video recordings of moving vehicles at traffic light junctions on the urban roadway. The average percentage for speed estimation error is 20.86%. The average percentage for accuracy obtained is 79.14%, and the overall average precision of 0.08
Value-based modeling and simulation for sustainable ICT4D
Get PDFICT4D is an acronym for information and communication technologies for development. This acronym represents the many initiatives that are now being carried out in underdeveloped countries all around the globe. These technology-oriented initiatives aim to promote the growth and prosperity of the regions they serve. People from the local community often volunteer their time to serve as program managers when NGOs or governments give funding. The term "sustainability" describes an information and communications technology's capacity to carry out a development project even after the initial funding has run out. The e3value technique will be broken down in this article so that we may better understand how it works. Using various tools for value-based modeling, this method is utilized to create an e-Business model. However, one of the main problems with e-business technologies is that they often get insufficient data to be effective. This study aims to examine the procedures that must be followed to learn more about an e-Business concept and mold it into a form that would allow for its implementation in a technologically and commercially practical manner. In order to gather information on the Long Lamai community's economic activities, a survey is being used in this research. The net value sheet for each of their businesses is then stimulated utilizing the data using E3 value. Initial findings indicate that 5 out of 13 enterprises' net incomes are negative, and each of those five businesses has its roots in a local enterprise. However, the figure does not consider labor costs, taxes, or long-term investments made with government money. We continued the stimulation with sensitivity analysis and scenario-based analysis, which results in total revenue of 1,410,000 for the five enterprises in the fourth year after accounting for taxes and fees in the sensitivity analysis. This opens the door for further research into how discrepancies in wealth and income may arise in the simulation of an economy due to the right use of ICT for long-term development and prosperity. Profitability sheets may be created if certain business model assumptions are specified, such as the monetary value of the commodities produced, distributed, and consumed. These may be used to decide if the project can become profitable for all parties concerned
Classification of Industrial Relations Dispute Court Verdict Document with XGBoost and Bidirectional LSTM
Get PDFIndustrial relations disputes (Perselisihan Hubungan Industrial (PHI)) are essential to examine because these disputes represent unbalanced bargaining positions between workers and corporations. On the other hand, there are many PHI documents, so they need to be classified and distinguished from other types of other decisions for other types of civil cases. PHI decisions document can be accessed openly from a special directory of civil courts. This ruling has similarities with other decisions regarding consumer protection or bankruptcy. This study used 450 documents consisting of 255 PHI court decisions and 255 non-PHI court decisions. This study takes the case as a classified part. We use several feature extractions and three methods: Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), and Bidirectional Long Short-Term Memory (Bi-LSTM). For SVM and XGBoost classifier, we utilize Frequency-inverse document frequency (TF-IDF). Another classifier needs word embedding Glove Wikipedia Indonesian with a dimension size of 50. Various experiments conducted found that the best classification results used Bi-LSTM with Gloves. This classification has 100% accuracy without overfitting. We found the second result using XGBoost with parameters optimized using random search, while the lowest accuracy results were obtained using the SVM method. The accuracy of the classification results in this study can impact the availability and quality of open legal knowledge that can be utilized by society and for future research
LoRaWAN for Smart Street Lighting Solution in Pangandaran Regency
Get PDFSmart street lighting is a key application in smart cities, enabling the monitoring and control of street lamps through internet connectivity. LoRa/LoRaWAN, an IoT technology, offers advantages such as low power consumption, cost-effectiveness, and a wide area network. With its extensive coverage of up to 15 kilometers and easy deployment, LoRa has become a favored connectivity option for IoT use cases. This study explores the utilization of LoRaWAN in Pangandaran, a regency in the West Java province of Indonesia. Implementing LoRaWAN in this context has resulted in several benefits, including the ability to monitor and control street lighting in specific areas of Pangandaran and real-time recording of energy consumption. The primary objective of this research is to estimate the number of LoRaWAN gateways required to support smart street lighting in Pangandaran. Two methods are employed: coverage calculation using the free space loss approach and capacity calculation. The coverage calculation suggests a requirement of 34 gateways, whereas the capacity calculation indicates that only two gateways are needed. Based on these findings, it can be inferred that, theoretically, a maximum of 34 gateways would be necessary for smart street lighting in the Pangandaran area. However, further research, including driving tests, is recommended to validate these results for future implementation. This study provides insights into the practical application of LoRaWAN technology in smart street lighting, specifically in Pangandaran. The findings contribute to optimizing infrastructure and resource allocation, ultimately enhancing the efficiency and effectiveness of urban lighting systems.
Prediction of Cross-Platform and Native Apps Technology Opportunities for Beginner Developers Using C 4.5 and Naive Bayes Algorithms
Get PDFThe competition between native and cross-platform app development makes application development simpler, safer, and more scalable. However, developers must have sufficient fundamentals, and the industry must conduct good research to shorten development time and minimize expenses. In order to solve these problems, this study made a prediction that discusses the technology that has a chance to survive in the industry so as not to be left behind in technology. Using Naïve Bayes and C 4.5 algorithms into a dataset with nine programming languages related to mobile app development. Results obtained in This research show Dart as a programming language that supports cross-platform frameworks and Kotlin as a programming language that supports native app frameworks is a technology that would have the opportunity in the future with an accuracy level above 90% with Naïve Bayes and C 4.5 algorithms. These results are obtained by testing an algorithm model using MAPE, consistent dataset sharing, and careful data processing. This research Can help entry-level developers learn and deepen the fundamentals of technology and can add knowledge to the industry in choosing a technology
Stock Price Time Series Data Forecasting Using the Light Gradient Boosting Machine (LightGBM) Model
Get PDFIn the world of stock investment, one of the things that commonly happens is stock price fluctuations or the ups and downs of stock prices. As a result of these fluctuations, many novice investors are afraid to play stocks. However, on the other hand, stocks are a type of investment that can be relied upon during disasters or economic turmoil, such as in 2019, namely the Covid-19 pandemic. For stock price fluctuations to be estimated by investors, it is necessary to carry out a forecasting activity. This study builds stock price forecasting using the Light Gradient Boosting Machine (LightGBM) algorithm, which has high accuracy and efficiency. To forecast stock price time series, the model used is the LightGBM ensemble. At the same time, they were optimizing the determination of hyperparameters using Grid Search Cross Validation (GSCV). This study will also compare the LGBM algorithm with other algorithms to see which model is optimal in forecasting price stock data. In this study, the test used the RMSE metric by comparing the original data (testing data) with the predicted results. The experimental results show that the LightGBM model can compete with and outperform boosting-based forecasting models like XGBoost, AdaBoost, and CatBoost. In comparing forecasting models, the same dataset is used so that the results are accurate, and the comparisons are equivalent. In future research, paying attention to the data during pre-processing is necessary because it has many outliers. In addition, it is necessary to include exogenous variables and external variables, which are determined to involve many parties
Comparison of the Packet Wavelet Transform Method for Medical Image Compression
Get PDFMedical images are often used for educational, analytical, and medical diagnostic purposes. Medical image data requires large amounts of storage on computers. Three types of codecs, namely Haar, Daubechies, and Biorthogonal, were used in this study. This study aims to find the best wavelet method of the three tested wavelet methods (Haar, Daubechies, and Biorthogonal). This study uses medical images representing USG and CT-scan images as testing data. The first test is carried out by comparing the threshold ratio. Three threshold values are used, namely 30, 40, and 50. The second test looks for PSNR values with different thresholds. The third test looks for a comparison of the rate (image size) to the PSSR value. The final test is to find each medical image's compression and decompression times. The first compression ratio test results on both medical images showed that CT scan images on Haar and Biorthogonal wavelets were the best, with an average compression ratio of 40.76% and a PSNR of 33.77. The PSNR obtained is also getting more significant for testing with a larger image size. The average compression time is 0.52 seconds, and the decompression time is 2.27 seconds. Based on the test results, this study recommends that the Daubechies wavelet method is very good for compression, which is 0.51 seconds, and the Biorthogonal wavelet method is very good for medical image decompression, which is 1.69 seconds