Article Per Year (5 Year)
p-Index From 2021 - 2026
1.532
P-Index
This Author published in this journals
All Journal IAES International Journal of Artificial Intelligence (IJ-AI) MEDIA KONSERVASI Mechatronics, Electrical Power, and Vehicular Technology Bulletin of Electrical Engineering and Informatics International Journal of Advances in Intelligent Informatics Register: Jurnal Ilmiah Teknologi Sistem Informasi EMITTER International Journal of Engineering Technology Buletin Ilmiah Sarjana Teknik Elektro Applied Engineering and Technology
Kohei Arai
Graduate School of Science and Engineering, Saga University, Saga City, Japan
Automotive Engineering Civil Engineering, Building, Construction & Architecture Computer Science & IT Earth & Planetary Sciences Education Electrical & Electronics Engineering Engineering Industrial & Manufacturing Engineering Materials Science & Nanotechnology

Published : 13 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 13 Documents
Search

 1   2 
Optimized Yolov8 to identify people with disabilities Wulanningrum, Resty; Handayani, Anik Nur; Herwanto, Heru Wahyu; Arai, Kohei
International Journal of Advances in Intelligent Informatics Vol 11, No 4 (2025): November 2025
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26555/ijain.v11i4.1977

Abstract

This research aims to develop an object detection model that can distinguish between the gait of people with and without disabilities with high accuracy. Object detection is currently designed to detect people and is used in both normal and gender-based gait recognition. Gait recognition, if further examined, encompasses recognition of both non-disabled and disabled individuals. Every day, people walk like most, but people with disabilities have different gaits from those of normal people. Some use walking aids, whereas others walk without them. YOLOv8 is a platform for detecting people. This research proposes an object detection for normal people and people with disabilities, both those who use assistive devices and those who do not. The dataset used is Disabled gait, comprising 6500 images, and will be divided into 3 data splits: 70% for training, 20% for validation, and 10% for testing. Model evaluation is based on precision, recall, mAP50, and mAP50-90. The test results for three classifications, namely assistive, non-assistive, and normal, show the highest value in the assistive class with an mAP50 value of 0.98 and an mAP50-95 value of 0.996. This study advances gait recognition by extending object detection to accurately differentiate normal and disabled walking patterns, including both assistive and non-assistive gaits, thereby enriching inclusive human-movement analysis. Beyond computer vision, the findings benefit healthcare, rehabilitation, and smart surveillance systems by enabling more accurate mobility assessment and accessibility-aware applications.
Centronit: Initial Centroid Designation Algorithm for K-Means Clustering Barakbah, Ali Ridho; Arai, Kohei
EMITTER International Journal of Engineering Technology Vol 2 No 1 (2014)
Publisher : Politeknik Elektronika Negeri Surabaya (PENS)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24003/emitter.v2i1.17

Abstract

Clustering performance of the K-means highly depends on the correctness of initial centroids. Usually initial centroids for the K- means clustering are determined randomly so that the determined initial centers may cause to reach the nearest local minima, not the global optimum. In this paper, we propose an algorithm, called as Centronit, for designation of initial centroidoptimization of K-means clustering. The proposed algorithm is based on the calculation of the average distance of the nearest data inside region of the minimum distance. The initial centroids can be designated by the lowest average distance of each data. The minimum distance is set by calculating the average distance between the data. This method is also robust from outliers of data. The experimental results show effectiveness of the proposed method to improve the clustering results with the K-means clustering.Keywords: K-means clustering, initial centroids, Kmeansoptimization.
Adaptive Feature Selection using Fisher-Based Supervised Hill Climbing for Dysgraphia Handwriting Classification Kirana, Kartika Candra; Handayani, Anik Nur; Eva, Nur; Wibawa, Aji Prasetya; Hidayat, Wahyu Nur; Arai, Kohei
Buletin Ilmiah Sarjana Teknik Elektro Vol. 8 No. 2 (2026): April
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/biste.v8i2.14983

Abstract

Dysgraphia features selection remains a challenge. Fisher’s criterion excels at highlighting the discriminative features of dysgraphia but lacks guidance for choosing the optimal number of features. Whereas Hill Climbing shows robust feature selection but often gets trapped in local optima. This study aims to avoid the Hill Climbing trap in local optima when selecting the best dysgraphia feature. Thus, the Fisher-Based Supervised Hill Climbing (FSHC) method is introduced. The contribution of this study is an optimized machine-learning-guided hill-climbing method that uses a classifier on a validation set as the objective function. A plateau mechanism also guided Hill Climbing exploration, not by a single Fisher point but by the neighboring subsets. The dataset used contains the graphomotor slant line task from 119 children aged 8-15 years (47.5% diagnosed with dysgraphia), with 10000 to 50000 data points per user. It is organized into kinematic, spatial, dynamic, and temporal features, yielding 117 sub-features. A stratified 5-fold cross-validation is set for training and testing, reaching 21 features. Comparative test—Linear SVM, SVM RBF, Sigmoid SVM, Polynomial SVM, Random Forest, AdaBoost, KNN, Decision Tree, Gradient Boosting, Gaussian Naive Bayes, and Gaussian Classifier—showed that linear SVM achieves the best performance with a weighted average precision, recall, and F1 score of 0.93. Linear SVM also outperformed the three approaches: no feature selection, the traditional Fisher, and machine-learning-based feature selection (weighted KNN and SVM). It can be concluded that the proposed method is more robust than the state of the art by highlighting key points for avoiding overfitting.
Co-Authors Aji Prasetya Wibawa Ali Ridho Barakbah Andrie Asmara, Rosa Anik Nur Handayani Ariyanta, Nadindra Dwi Burhanuddin, Mohd Aboobaider Damayanti, Farradila Ayu Daud, Nathan Dimas Wahyu Wibowo Eva, Nur Faqih, Kamil Heru Wahyu Herwanto Hitipeuw, Emanuel Indriati Indriati Jevri Tri Ardiansah Kartika Candra Kirana Kurnianingtyas, Diva Marji Marji Mustika Mentari Nugraha, Agil Zaidan Nur Hidayat, Wahyu Priharta, Ari Primadi, Wahyu Rahmat Samudra Anugrah, Muhammad Resty Wulanningrum Rosa Andrie Asmara Sakti Oktafiansyah, Haidar Salsabila, Reni Fatrisna