Claim Missing Document
Check
Articles

Found 32 Documents
Search

Optimizing the viola-jones algorithm for robust face recognition in variable lighting and orientation conditions Gunawan, Gunawan; Aisyah, Nur; Santoso, Nugroho Adhi
Jurnal Mantik Vol. 8 No. 1 (2024): May: Manajemen, Teknologi Informatika dan Komunikasi (Mantik)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mantik.v8i1.5220

Abstract

Facial recognition is a critical technology in digital security, driven by significant advances in computer vision. This research focuses on optimizing the Viola-Jones algorithm to improve the accuracy and speed of face detection by adjusting parameters and integrating more sophisticated image processing techniques. Facing challenges such as suboptimal lighting and variations in face orientation, the study adopted a rigorous experimental design, in-depth quantitative analysis, and robust model validation. Of the ten facial images collected, all were intensively processed using Haar-like features to identify significant patterns and adjust algorithm parameters in Python. This optimization process increased performance from 7 identified faces to 9 post-optimization identified faces and a substantial decrease in detection time from 0.0065 seconds to 0.0017 seconds per image. The comprehensive evaluation showed an increase in accuracy from 70% to 90%, recall from 70.0% to 90.0%, Precision remained constant at 100.0%, and F1-score from 82.35% to 94.74%. These results show that the optimization has increased the algorithm's sensitivity to changes in light intensity and face orientation and improved the effectiveness of facial recognition systems in complex and dynamic security scenarios while providing concrete evidence of the benefits of using Haar-like features in the Viola-Jones algorithm
Application of ant colony algorithm to optimize waste transport distribution routes in Tegal Gunawan, Gunawan; Handayani, Sri; Anandianskha, Sawaviyya
Jurnal Mantik Vol. 8 No. 1 (2024): May: Manajemen, Teknologi Informatika dan Komunikasi (Mantik)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mantik.v8i1.5223

Abstract

Effective and efficient waste management is an essential challenge in developing cities like Tegal City. Optimizing waste transport routes can reduce operational costs and environmental impact. This study aims to implement the Ant Colony Algorithm (ACO) to optimize waste distribution routes in Tegal City. This method was chosen for its proven ability to solve route optimization problems. This study developed a model for the simulation and analysis of waste transportation routes using actual location data from the Integrated Waste Treatment Site (TPST) to the landfill (TPA). The results showed that the implementation of ACO reduced the total mileage from 27.50 km to 21.05 km, a significant reduction that shows the algorithm's efficiency in determining the optimal route. The conclusion of this study confirms that ACO can be effectively used to improve waste transportation operations
Application of artificial neural network with optimization of genetic algorithms for weather prediction Gunawan, Gunawan; Miftakhudin, Muhammad; Arif, Zaenul
Jurnal Mantik Vol. 8 No. 1 (2024): May: Manajemen, Teknologi Informatika dan Komunikasi (Mantik)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mantik.v8i1.5225

Abstract

This research integrates Artificial Neural Network (ANN) with Genetic Algorithm Optimization (GA) to improve the accuracy of weather prediction. This method utilizes ANN-optimized GA, creating a model that can adapt to the dynamics of weather patterns. Using a dataset that includes meteorological variables such as temperature, humidity, and precipitation from January 1, 2023, to October 28, 2023, the model was tested for its ability to predict weather conditions accurately. The process begins with data preprocessing, ANN training, and GA optimisation. The evaluation showed that the optimized model was able to reduce the Mean Absolute Error (MAE) from 1.6865 to 0.8701, the Mean Absolute Percentage Error (MAPE) from 5.9864 to 3.1408, and the Root Mean Squared Error (RMSE) from 2.253 to 1.039, signalling a significant improvement in prediction accuracy and efficiency. This research confirms the potential of ANN and GA integration in improving weather prediction, providing new insights for developing more accurate and reliable prediction models for various applications, from agriculture to disaster management.
Application of the analytic network process method in the selection of raw material suppliers for yarn Gunawan, Gunawan; Hafid Subechi, Fadlan; Arif, Zaenul
Jurnal Mantik Vol. 8 No. 1 (2024): May: Manajemen, Teknologi Informatika dan Komunikasi (Mantik)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mantik.v8i1.5227

Abstract

This study applies the Analytic Network Process (ANP) Method for selecting raw material suppliers for yarn, a crucial factor in boosting production efficiency and quality within the textile industry. The research aims to develop and validate a decision-making model that enhances supplier selection by integrating ANP with rigorous quantitative analyses. The methodology incorporates a series of experiments, thorough examination of historical data, and robust model validation processes to confirm the accuracy and dependability of the findings. The results demonstrate significant improvements in the precision of supplier selection, underscored by a high Pearson correlation coefficient of 0.89. This validates the model's effectiveness and reliability, suggesting that the developed framework not only supports data-driven and objective decision-making in the textile industry but also has potential applications in other sectors to enhance operational efficiency and sustainability
Application of the rule-based system method to determine the type of crops based on altitude and rainfall Gunawan, Gunawan; Firmansyah, Akhmad Lutfi; Santoso, Bayu Aji
Jurnal Mantik Vol. 8 No. 1 (2024): May: Manajemen, Teknologi Informatika dan Komunikasi (Mantik)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mantik.v8i1.5234

Abstract

Applying the rule-based system method to determine the type of agricultural crop based on altitude and rainfall is essential in increasing productivity and efficiency in modern agriculture. This study aims to develop and implement a rules-based system to recommend suitable plant types by analyzing altitude and rainfall data in the Tegal District. The research method includes experimental design, quantitative analysis, and model validation using data from the Central Bureau of Statistics and various other internet sources, covering January 1 to December 31, 2023. The results showed that this rule-based system effectively provides accurate recommendations with an average accuracy rate of 85% and an error rate of 15%. This system helps farmers make informed decisions about crop selection, reducing crop failure risk and contributing to sustainable agricultural practices. Future research suggests integrating real-time weather prediction technology and additional environmental variables to improve the precision of recommendations and expand the applicability of these systems to other areas with similar characteristics
Expert system for diagnosing diseases in corn plants using the navies bayes method Gunawan, Gunawan; Sya’bani, Adhita Zulfa; Anandianshka, Sawaviyya
Jurnal Mantik Vol. 8 No. 1 (2024): May: Manajemen, Teknologi Informatika dan Komunikasi (Mantik)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mantik.v8i1.5265

Abstract

This research introduces an expert system using the Naive Bayes method to diagnose corn plant diseases, aiming to provide an automated, accurate, and scalable diagnostic tool. Traditional methods are often inefficient and error-prone, relying on expert knowledge and manual inspection. This study employs a quantitative approach, incorporating experimental design, data analysis, and model validation. Data on humidity, temperature, and soil conditions were collected from agricultural research centers and online databases. After preprocessing, key variables influencing disease occurrence were selected. The Naive Bayes model was optimized using cross-validation and implemented in Python, achieving an average accuracy of 92%. The model's performance, evaluated through accuracy, precision, recall, and F1-score, demonstrated the effective distinction between similar symptoms—the system's simplicity and computational efficiency suit resource-constrained environments like rural farms. By combining visual symptoms and environmental factors, the system minimizes dependency on expert knowledge, offering a comprehensive and scalable solution for disease management in agriculture
Application of k-nearest neighbors method for detection of beef authenticity based on beef image gunawan, Gunawan; Moonap, Dinar Auranisa; Fadhilah, Nurul
Jurnal Mantik Vol. 8 No. 1 (2024): May: Manajemen, Teknologi Informatika dan Komunikasi (Mantik)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/mantik.v8i1.5281

Abstract

Beef authenticity detection is a significant concern in today's food industry. This study proposes the K-Nearest Neighbors (K-NN) method based on the extraction of the Histogram of Oriented Gradients (HOG) feature to detect the authenticity of beef based on images. A dataset of 40 images of real and fake beef was collected and aggregated into 240 images to increase the variety of data. The imagery is changed to grayscale, and the HOG feature is extracted to capture texture and shape information. The K-NN model is built with optimized parameters using Grid Search and cross-validation techniques. The model was evaluated by measuring accuracy, precision, recall, and F1-score on the test data. The results show that the K-NN model with HOG feature extraction can achieve an accuracy of 80.56%,  precision of 87.10%, recall of 72.97%, and F1-score of 72.97% in classifying real and fake beef. These findings confirm the effectiveness of the proposed method for the rapid and accurate detection of beef authenticity. This research contributes to developing image-based food authenticity detection methods that can be applied to increase consumer confidence in the food industry
Application of computer vision techniques to detect diseases and pests of chili plants Nurokhman, Akhmad; Surorejo, Sarif; Kurniawan, Rifki Dwi; Gunawan, Gunawan
Journal of Intelligent Decision Support System (IDSS) Vol 7 No 1 (2024): March: Intelligent Decision Support System (IDSS)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/idss.v7i1.201

Abstract

This research aims to develop a disease and pest detection system in chili plants using computer vision techniques. In this study, deep learning methods, especially Convolutional Neural Networks (CNN), were applied to identify and classify various types of diseases and pests that often attack chili plants. The data used included images of chili leaves infected with various diseases and pests, which were then trained in CNN models to recognize certain patterns that indicate the presence of infection. The results showed that the developed system was able to detect and classify diseases and pests in chili plants with a very high degree of accuracy. The novelty of this research lies in the use of computer vision techniques combined with sophisticated deep learning algorithms to automatically detect diseases and pests, which were previously done manually by farmers or agricultural experts. These findings make an important contribution to improving efficiency and effectiveness in chili crop health management, offering innovative solutions to support agricultural sustainability through the use of advanced technology.
Application of computer vision for face recognition using viola jones algorithm method Riyadi, Fajar Sugeng; Gunawan, Gunawan; Arif, Zaenul
Journal of Intelligent Decision Support System (IDSS) Vol 7 No 1 (2024): March: Intelligent Decision Support System (IDSS)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/idss.v7i1.204

Abstract

This research aims to develop a facial recognition system using computer vision technology by applying the Viola-Jones algorithm method. The main focus of this research is to improve accuracy and efficiency in face identification under various lighting conditions and face orientations. The Viola-Jones algorithm, known for its real-time object detection, was chosen for its efficiency in quickly identifying critical facial features. Through testing of various face datasets, the results showed that the system developed was able to recognize faces with a high level of accuracy, even in conditions of non-optimal lighting and various facial poses. The novelty of this research lies in the optimization of the parameters of the Viola-Jones algorithm to improve facial recognition performance, as well as its application in challenging dynamic environments. These findings make a significant contribution to the field of computer vision and facial recognition, offering more effective and efficient solutions for security and surveillance applications, as well as interactive applications that require fast and accurate facial identification.
Classification of fresh chicken meat and tainted chicken meat using naive bayes classifier algorithm Zain, Ahmad Muzakky; Ali Murtopo, Aang; Fadila, Nurul; Gunawan, Gunawan
Journal of Intelligent Decision Support System (IDSS) Vol 7 No 1 (2024): March: Intelligent Decision Support System (IDSS)
Publisher : Institute of Computer Science (IOCS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35335/idss.v7i1.212

Abstract

This research discusses the classification of fresh and tainted chicken meat using the Naive Bayes Classifier (NBC) algorithm based on Gray Level Co-occurrence Matrix (GLCM) feature extraction, with the aim of developing an efficient and accurate classification method. This research aims to utilize image processing and machine learning technologies to distinguish fresh chicken meat from tainted ones, which is crucial for the food industry. The research methodology involved the use of GLCM for texture feature extraction from chicken meat images, with the implementation of the NBC model through RapidMiner, offering an intuitive and efficient approach. The results showed the success of the model in achieving 80% accuracy, with an average precision of 81.25%, recall of 80%, and F1-score of 80.62%, confirming its ability in chicken meat classification. The integration of GLCM and RapidMiner in the application of NBC not only improves accuracy and objectivity in chicken meat classification but also provides a foundation for the wider application of machine learning techniques in ensuring food safety and consumer satisfaction