Garuda - Garba Rujukan Digital

Comparative Analysis of Keypoint Detection Performance in SIFT Implementations on Small-Scale Image Datasets Arif Rahman; Suprihatin; Imam Riadi; Tawar; Furizal
Journal of Innovation Information Technology and Application (JINITA) Vol 6 No 2 (2024): JINITA, December 2024
Publisher : Politeknik Negeri Cilacap

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35970/jinita.v6i2.2399

Scale-invariant feature transform (SIFT) is widely used as an image local feature extraction method because of its invariance to rotation, scale, and illumination change. SIFT has been implemented in different program libraries. However, studies that analyze the performance of SIFT implementations have not been conducted. This study examines the keypoint extraction of three well-known SIFT libraries, i.e., David Lowe's implementation, OpenSIFT, and vlSIFT in vlfeat. Performance analysis was conducted on multiclass small-scale image datasets to capture the sensitivity of keypoint detection. Although libraries are based on the same algorithm, their performance differs slightly. Regarding execution time and the average number of keypoints detected in each image, vlSIFT outperforms David Lowe’s library and OpenSIFT.

A Bibliometric Analysis of Natural Language Processing and Classification: Trends, Impact, and Future Directions Setiawan Ardi Wijaya; Rahmad Gunawan; Rangga Alif Faresta; Asno Azzawagama Firdaus; Gabriel Diemesor; Furizal
Scientific Journal of Engineering Research Vol. 1 No. 1 (2025): March
Publisher : PT. Teknologi Futuristik Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.64539/sjer.v1i1.2025.6

This study presents a bibliometric analysis of Natural Language Processing (NLP) and classification research, examining trends, impacts, and future directions. NLP, a key field in artificial intelligence, focuses on enabling computers to process and understand human language through tasks such as text classification, sentiment analysis, and speech recognition. Classification plays a crucial role in organizing textual data, facilitating applications like spam detection and content recommendation. The research employs bibliometric analysis to evaluate publication trends, citation networks, and emerging themes from 1992 to 2025. Using data retrieved from Scopus, descriptive statistical analysis and bibliometric mapping with VOSviewer reveal key contributors, influential publications, and subject area distributions. Findings indicate a significant rise in NLP research, with deep learning models, particularly transformers, driving advancements in the field. The study highlights dominant research areas, including computer science, engineering, and medicine, and identifies leading countries in NLP research, such as the United States, China, and India. Additionally, ethical concerns, including bias and fairness in NLP applications, are discussed as critical challenges for future research. The insights derived from this analysis provide valuable guidance for researchers and policymakers in shaping the next phase of NLP development.

Early Detection of Brain Tumors: Performance Evaluation of AlexNet and GoogleNet on Different Medical Image Resolutions Muis, Alwas; Rustiawan, Angga; Oyeyemi, Babatunde Bamidele; Syukur, Abdul; Furizal
Scientific Journal of Engineering Research Vol. 1 No. 3 (2025): September
Publisher : PT. Teknologi Futuristik Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.64539/sjer.v1i3.2025.10

Early detection of brain tumors through medical imaging is crucial to improving treatment success rates. This study aims to classify brain tumors using two deep learning models, AlexNet and GoogleNet, by testing three image sizes. The dataset used consists of four classes: glioma, no tumor, meningioma, and pituitary. The test results show that the AlexNet model achieves the best accuracy of 98% at a resolution of 150x150, while GoogleNet shows stable performance with the highest accuracy of 96% at both 150x150 and 200x200 resolutions. The medium resolution (150x150) proves to be optimal for both models, providing the best balance between visual information and processing efficiency. This study highlights the potential use of AlexNet and GoogleNet in brain tumor classification, with opportunities for performance improvement through further development, such as ensemble techniques and the use of a larger dataset.

Classification for Waste Image in Convolutional Neural Network Using Morph-HSV Color Model Fahmi, Miftahuddin; Yudhana, Anton; Sunardi; Abdel-Nasser Sharkawy; Furizal
Scientific Journal of Engineering Research Vol. 1 No. 1 (2025): March
Publisher : PT. Teknologi Futuristik Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.64539/sjer.v1i1.2025.12

Waste management is essential in preserving nature to be cleaner and more well-maintained. Waste management runs slower than the speed of waste accumulation. One reason is slow waste sorting. This problem can be overcome by building a learning machine that can sort the types of waste. The type of waste often separated in the first sorting is waste based on its type, namely organic and inorganic. The classification model used is the CNN with image processing Morph-HSV color model. The data obtained from Kaggle is collected and processed using Python. The processed image is trained using a CNN classification model. The results of this study are an accuracy of 99.58% and a loss of 1.57%. With this research, it is hoped that it can accelerate waste sorting performance using the most efficient ML based on image processing and its classification model.

Comparative Analysis of Keypoint Detection Performance in SIFT Implementations on Small-Scale Image Datasets Arif Rahman; Suprihatin; Imam Riadi; Tawar; Furizal
Journal of Innovation Information Technology and Application (JINITA) Vol 6 No 2 (2024): JINITA, December 2024
Publisher : Politeknik Negeri Cilacap

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35970/jinita.v6i2.2399

Scale-invariant feature transform (SIFT) is widely used as an image local feature extraction method because of its invariance to rotation, scale, and illumination change. SIFT has been implemented in different program libraries. However, studies that analyze the performance of SIFT implementations have not been conducted. This study examines the keypoint extraction of three well-known SIFT libraries, i.e., David Lowe's implementation, OpenSIFT, and vlSIFT in vlfeat. Performance analysis was conducted on multiclass small-scale image datasets to capture the sensitivity of keypoint detection. Although libraries are based on the same algorithm, their performance differs slightly. Regarding execution time and the average number of keypoints detected in each image, vlSIFT outperforms David Loweâ€™s library and OpenSIFT.

Ball Detection System for a Soccer on Wheeled Robot Using the MobileNetV2 SSD Method Puriyanto, Riky D.; Yunandha, Isro D.; Maghfiroh, Hari; Ma'arif, Alfian; Furizal; Suwarno, Iswanto
Emerging Science Journal Vol. 9 No. 5 (2025): October
Publisher : Ital Publication

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.28991/ESJ-2025-09-05-028

This paper discusses the research on the use of Artificial Intelligence in autonomous robot object identification. The specific focus of this research is on a wheeled soccer playing robot. The goal is to recognize a ball as an object using the Single Shot MultiBox Detector MobileNetV2 model. This system has multi-vision inputs such as distance measurements and angle values for object detection. This methodology is based on deep learning with the TensorFlow Object Detection API with the MobileNetV2 SSD model. This model is trained with a dataset of 3707 ball images over 617 thousand steps on Google Collaboratory. It was found that the average measurement error of the ball object is 6.58% for the distance when viewed through the robot's front camera. In addition, the omnidirectional camera is able to detect the ball object and angle values from the front of the robot. What makes this research different is the use of distance and angle measurements for detection and the omnidirectional camera for system performance in dynamic environments. This research aims to address the improvement of AI-based object detection systems for autonomous robotics in the context of real-world use cases.

DC Motor Angular Speed Controller Using an Embedded Microcontroller-Based PID Controller Ma'arif, Alfian; Nugraha, Ikhwan; Maghfiroh, Hari; Furizal; Suwarno, Iswanto
Emerging Science Journal Vol. 9 No. 6 (2025): December
Publisher : Ital Publication

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.28991/ESJ-2025-09-06-03

This research presents the implementation of a Proportional Integral Derivative (PID) controller to control the angular speed of a Direct Current (DC) motor using an embedded system (microcontroller). The system’s hardware consists of an Arduino microcontroller, a DC motor with an encoder sensor, a driver motor, and a power supply. Proportional control regulates the response proportionally to the calculated error, while integral control manages the cumulative error over time, and derivative control responds to the rate of change of the error, preventing overshoot. With a proper combination, PID control achieves stability, speeds up response, and reduces overshoot, improving overall system performance. Based on experimental data, the DC motor angular speed control system using PID control achieves the best results, in which the parameter values are Kp=1; Ki=0.3; and Kd=0.6. The augmented system responded with 0.0890 seconds of the rise time, 11.772 seconds of settling time, and 0.12 seconds of the peak time, with an overshoot of less than 10% (7%).

Machine Learning 5.0 In-depth Analysis Trends in Classification Dianda Rifaldi; Tri Stiyo Famuji; Setiawan Ardi Wijaya; Ahmed Jaber Abougarair; Phichitphon Chotikunnan; Alfian Ma'arif; Furizal
Scientific Journal of Computer Science Vol. 1 No. 1 (2025): June
Publisher : PT. Teknologi Futuristik Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.64539/sjcs.v1i1.2025.18

In the era of Technology 5.0 Machine Learning continues to show significant advancements across various sectors. This study aims to examine the latest trends in Machine Learning classification, focusing on four key approaches Explainable Artificial Intelligence, Federated Learning, Transfer Learning, and Generative Adversarial Networks. The methodology involves a comprehensive literature review of research in Asia and experimentation with related datasets. The findings indicate that Explainable Artificial Intelligence enhances transparency and accuracy in data classification, Federated Learning enables decentralized learning while safeguarding data privacy, Transfer Learning improves accuracy with small datasets, and Generative Adversarial Networks aids in data augmentation for better model training. In conclusion, these techniques not only enhance the efficiency and accuracy of classification but also open up new opportunities for innovation in various fields, including healthcare, transportation, and cybersecurity.

Title

Found 8 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 8 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 8 Documents
Search