Claim Missing Document
Check
Articles

Found 2 Documents
Search
Journal : Indonesian Journal of Electrical Engineering and Computer Science

Efficient object detection for augmented reality based english learning with YOLOv8 optimization Putra, Arya Krisna; Tambunan, Fiqri Ramadhan; Ndruru, Samson; Chowanda, Andry
Indonesian Journal of Electrical Engineering and Computer Science Vol 39, No 2: August 2025
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijeecs.v39.i2.pp1189-1197

Abstract

This study develops a mobile-based augmented reality (AR) application with machine learning for elementary school students to enhance basic English vocabulary learning. The application integrates an optimized YOLOv8 object detection model, designed to recognize 20 common classroom objects in real-time. The model optimization involves replacing standard Conv layers with GhostConv and the C2f block with the C2fCIB block that has significantly improved computational efficiency. Evaluation results show the optimized model reduces the parameters by 22.003% and decreases the file size from 6.2 MB to 4.9 MB. The model performance improved by achieving precision of 83.7%, recall of 73.5% and a mean Average Precision (mAP) of 81.4%. The model was integrated into the Unity platform via the Barracuda library, enabling real-time detection and interactive display of 3D objects. This aplication also complete with English text, translations, example sentences also audio pronunciation. 3D objects representing classroom vocabulary were specifically created to support AR-based learning. Performance testing on a Samsung A14 showed an improved frame rate of 6–12 FPS compared to the original model’s 5–10 FPS. These results demonstrate that the optimized YOLO model effectively integrates with AR technology, creating a more interactive and enjoyable vocabulary learning experience.
Real-time recognition of Indonesian sign language SIBI using CNN-SVM model combination Santika, Satriadi Putra; Benhard, Stefanus; Arifin, Yulyani; Chowanda, Andry
Indonesian Journal of Electrical Engineering and Computer Science Vol 39, No 2: August 2025
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijeecs.v39.i2.pp1198-1210

Abstract

Real-time Sistem Isyarat Bahasa Indonesia (SIBI) sign language recognition plays a crucial role in improving accessibility for individuals with hearing and speech impairments. Despite advancements in SIBI recognition research, challenges remain in ensuring model stability and accuracy in realtime settings, particularly in handling gesture variations and classification inconsistencies. This study addresses these challenges by developing a convolutional neural network-support vector machine (CNN-SVM) combination model, integrating MediaPipe for hand coordinate extraction, CNN for feature extraction, and SVM for classification. To improve generalization and prevent overfitting, data augmentation is applied to expand the dataset. The model's performance is further enhanced through hyperparameter optimization (HPO) and post-processing techniques such as multi-window majority voting (MWMV) and SymSpell. Experimental results show that the CNN-SVM model trained on augmented data with HPO achieves 91% testing accuracy, outperforming both standalone CNN and SVM models. Furthermore, MWMV improves recognition stability, while SymSpell enhances spelling errors, ensuring more meaningful outputs. The system is integrated with OpenCV for real-time recognition, but current deployment remains limited to local execution. Future work will focus on developing lightweight models for web-based and mobile applications, making the system more accessible and scalable.