Article Per Year (5 Year)
p-Index From 2020 - 2025
0.751
P-Index
This Author published in this journals
All Journal IAES International Journal of Artificial Intelligence (IJ-AI) JOIV : International Journal on Informatics Visualization
Athala, Varyl Hasbi
Unknown Affiliation
Computer Science & IT Engineering

Published : 4 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 4 Documents
Search

 1 
Enhancing Vision-Based Vehicle Detection and Counting Systems with the Darknet Algorithm and CNN Model Rangkuti, Abdul Haris; Athala, Varyl Hasbi
JOIV : International Journal on Informatics Visualization Vol 9, No 1 (2025)
Publisher : Society of Visual Informatics

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62527/joiv.9.1.2550

Abstract

This study focuses on developing an algorithm that accurately calculates the volume of vehicles passing through a busy crossroads in Indonesia using object recognition. The high density of vehicles and their proximity often pose a challenge when distinguishing between vehicle types using a camera. Therefore, the proposed algorithm is designed to assign a unique identity (ID) to each vehicle and other objects, such as pedestrians, ensuring that volume calculations are not repeated. The objective is to provide an equitable comparison of road density and the total number of detected vehicles, enabling the determination of whether the road is crowded. To accomplish this, the algorithm incorporates the Non-Max Suppression function, which displays bounding boxes around objects with confidence values and counts the objects within each box. Even when objects are nearby, the algorithm tracks them effectively, thanks to the support of the Darknet Algorithm. The main capabilities of this algorithm for improving vehicle detection include enhanced accuracy, speed, and generalization ability. Typically, it is used in conjunction with the You Only Look Once (YOLO) object detection framework. Five convolutional neural network models are tested to assess the algorithm's accuracy: YOLOv3, YOLOv4, CrResNext50, DenseNet201-YOLOv4, and YOLOv7-tiny. The training process utilizes the Darknet Algorithm. The best-performing models, YOLOv3 and YOLOv4, achieve exceptional accuracy and F1 scores of up to 99%. They are followed by CrResNext50 and DenseNet201-YOLOv4, which achieve accuracy rates of 92% and 98% and F1 scores of 94% and 98%, respectively. The YOLOv7-tiny model achieves an accuracy rate and F1 score of 86% and 88%, respectively. Overall, the results demonstrate the algorithm's success in accurately detecting and calculating the volume of vehicles and other objects in a busy intersection. This makes it a valuable tool for regional government decision-making.
Identification of Indonesian Traditional Foods Using Machine Learning and Supported by Segmentation Methods Rangkuti, Abdul Haris; Kerta, Johan Muliadi; Mogot, Roderik Yohanes; Athala, Varyl Hasbi
JOIV : International Journal on Informatics Visualization Vol 8, No 4 (2024)
Publisher : Society of Visual Informatics

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62527/joiv.8.4.2545

Abstract

Traditional food is essential in preserving cultural heritage and is a vital part of Indonesian cuisine. In this research, we implement a methodology to identify the traditional Indonesian food using machine learning algorithms supported by various segmentation methods. This research aims to provide an efficient and accurate approach to classifying traditional foods, which can contribute to promoting and preserving Indonesia's culinary heritage. To conduct this research, we conducted experiments on 34 types of conventional Indonesian food originating from various provinces in Indonesia. The analysis of food images involved several segmentation algorithms, including Sobel, Prewitt, Robert, Scharr, and Canny filters. After the segmentation process, we proceeded with feature extraction and classification using traditional machine learning algorithms such as the Random Forest algorithm, Decision Tree, and derivatives of the SVM algorithm. These algorithms aimed to recognize the 34 types of traditional food. After conducting several experiments, we found that Random Forest with Robert's segmentation method was the highest-performance algorithm. It produced extraordinarily accurate results on the test dataset, with an accuracy performance of 85.52%, recall of 84.63%, precision of 83.77%, and an f1 score of 82.49%. Additionally, the best-performing algorithms with execution time averaged less than 1 minute. Another experimental result showed that the Random Forest algorithm with the Canny operator achieved an accuracy of 81.51%, recall of 84.97%, precision of 86.8%, and an f1 score of 85.61% on the test dataset. Furthermore, the Random Forest algorithm with the Sobel operator achieved accuracy results of 78.4%, recall of 65.3%, precision of 62.3%, and an f1 score of 63.71%.  In the SVM algorithms derivative, the Sigmoid SVM combined with the Scharr operator achieved the highest performance in its category across all classification metrics. In conclusion, this research offers valuable insights into classifying traditional Indonesian dishes using traditional machine learning algorithms. Simultaneously, this research aims to promote the appropriate and effective preservation and recognition of traditional Indonesian food.
Deep Learning Algorithms and Optimizers: Enhancing the Evaluation of Signature Authenticity Haris Rangkuti, Abdul; Tanuar, Evawaty; Kusuma, Verdiant Jonathan; Athala, Varyl Hasbi
JOIV : International Journal on Informatics Visualization Vol 9, No 2 (2025)
Publisher : Society of Visual Informatics

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62527/joiv.9.2.2625

Abstract

Given the rapid technological advancements, security has become an essential human need that must be addressed. For example, a signature, which serves as a unique identifier or mark on a document, is vital in verifying and legalizing its contents. This study aims to utilize image processing techniques to identify patterns in signature images. Generally, a signature is a handwritten depiction used to authorize a document, indicating that the signing party acknowledges and agrees to its contents. However, this practice exposes signatures to the risk of forgery by dishonest individuals. Therefore, it is crucial to implement a security system for identity recognition using a biometric system for verification and identification. Verification involves determining whether the signature belongs to a previously identified individual and assessing its authenticity. This study employs deep learning algorithms, enhanced by optimizer methods, to improve accuracy performance in signature recognition for authenticity. Additionally, classification methods such as Linear SVM and RbfSVM are utilized. Several experiments were conducted, with VGG16 paired with the Adam optimizer yielding the highest accuracy of 0.9923. This was closely followed by VGG19 with Adagrad and Xception with RMSprop, achieving an accuracy of 0.9915. The training and validation accuracy processes revealed that the CNN VGG19 and VGG16 models with the Adam optimizer consistently achieved an accuracy exceeding 99%. Based on these experimental findings, the accuracy for detecting genuine and fake signatures can be clearly distinguished with an accuracy of over 99%, demonstrating the validity of this approach.
Enhancing touchless smart locker systems through advanced facial recognition technology: a convolutional neural network model approach Rangkuti, Abdul Haris; Tanuar, Evawaty; Yapson, Febriant; Sijoatmodjo, Felix Octavio; Athala, Varyl Hasbi
IAES International Journal of Artificial Intelligence (IJ-AI) Vol 14, No 4: August 2025
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijai.v14.i4.pp3262-3273

Abstract

As the world recovers from COVID-19, demand for contactless systems is increasing, promising safety and convenience. Touchless technology, particularly public locker security systems that use facial recognition and hand detection, is advancing rapidly. The system minimizes physical contact, increasing user safety. It uses advanced models such as multi-task cascaded convolutional networks (MTCNN) and RetinaFace, FaceNet512, ArcFace, and visual geometry group (VGG)-Face for face detection and recognition, with a combination of RetinaFace, ArcFace, and L2 norm Euclidean or cosine as the most effective distance metric method, where the accuracy reaches 96 and 90%. 'Yourvault', an application demonstrating this efficient security feature, provides notifications for mask detection, facial authenticity and locker status, offering a solution to the problem of convenience and security of public spaces. Future research could investigate the impact of photo age on facial recognition accuracy, potentially making touchless systems more efficient. In general, the application of this technology is an important step towards a safer and more comfortable world after the pandemic. This model approach can be followed up with more optimal facial recognition.
Co-Authors Haris Rangkuti, Abdul Johan Muliadi Kerta, Johan Muliadi Kusuma, Verdiant Jonathan Mogot, Roderik Yohanes Rangkuti, Abdul Haris Sijoatmodjo, Felix Octavio Tanuar, Evawaty Yapson, Febriant