Claim Missing Document
Check
Articles

Found 3 Documents
Search
Journal : Jurnal Sistem informasi dan informatika (SIMIKA)

AI-BASED APPLICATION FOR INDONESIAN SIGN LANGUAGE DETECTION USING YOLOV8 Khansa, Devanna Alandra; Nurhaida, Ida
Jurnal Sistem Informasi dan Informatika (Simika) Vol. 8 No. 2 (2025): Jurnal Sistem Informasi dan Informatika (Simika)
Publisher : Program Studi Sistem Informasi, Universitas Banten Jaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47080/p8pmqy04

Abstract

Sign language is used by individuals with disabilities, particularly the deaf and those with speech impairments, as their primary means of communication. However, interaction between people with disabilities and the general public is often hampered by a lack of understanding of sign language. This study aims to develop an artificial intelligence-based application capable of detecting and classifying hand movements in Indonesian Sign Language (BISINDO) using the YOLOv8 algorithm. The YOLOv8 algorithm was chosen for its ability to detect and classify objects in real-time with high accuracy, even under varying lighting and background conditions. This is one of the first studies to implement YOLOv8 for real-time BISINDO detection integrated with a web interface. The dataset used includes 51 classes of hand movements with a total of 10,822 images that have undergone augmentation to increase data diversity. The development process involved data collection, pre-processing, annotation, model training, and integration with an interactive web interface. The resulting model demonstrated high performance, achieving mAP@50 of 96%, mAP@50-95 of 70%, and classification accuracy of 93.8% in the final evaluation. This application is intended to help the deaf community communicate more easily with the wider community. It can improve communication accessibility for individuals with hearing impairments in public and educational settings, as well as provide an innovative solution to support social inclusivity. Further testing and parameter optimization will be conducted to expand the detection coverage and improve the system's performance in the future.
AI-BASED FACIAL DE-IDENTIFICATION FOR CHILDREN'S DIGITAL PRIVACY Negara, Komang Putra Satria; Nurhaida, Ida
Jurnal Sistem Informasi dan Informatika (Simika) Vol. 8 No. 2 (2025): Jurnal Sistem Informasi dan Informatika (Simika)
Publisher : Program Studi Sistem Informasi, Universitas Banten Jaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47080/m11eck75

Abstract

Social media has become an inseparable part of daily life in today's digital era. Many parents frequently share photos of their children online, exposing them to risks related to privacy and security. This research addresses such issues by developing an Android-based facial de-identification application that utilises the YOLOv8 algorithm to protect minors' privacy. The methodology involves several stages: data collection, pre-processing, model training, and application development. The dataset includes over 2,889 images of children, which were augmented to enhance its size and diversity. YOLOv8, a state-of-the-art object detection algorithm, was trained with these images to achieve high precision and recall in identifying children's faces. The developed application integrates the YOLOv8 model within a user-friendly interface built with Flutter. Results indicate that YOLOv8 effectively detects children's faces with a high precision of 95%, accuracy of 88%, recall of 92%, and mAP50 of 0.977. While the model demonstrates strong performance on training data, there is room for improvement on unseen data. By leveraging YOLOv8 and providing an accessible mobile application, the work allows parents to protect their children's identities online. The application mitigates risks of unauthorised use and exploitation of children's images by enabling facial de-identification, thus promoting safer online practices for families.
FACE DETECTION AND ANTI-SPOOFING ON DESKTOP APPLICATIONS USING YOU ONLY LOOK ONCE Faisal, Fairo Mahaputranda; Nurhaida, Ida
Jurnal Sistem Informasi dan Informatika (Simika) Vol. 8 No. 2 (2025): Jurnal Sistem Informasi dan Informatika (Simika)
Publisher : Program Studi Sistem Informasi, Universitas Banten Jaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47080/6qntes73

Abstract

In the digital era, facial recognition systems have become increasingly vulnerable to spoofing attacks, as demonstrated by cases of identity theft using photos or smartphone screens. This study develops a real-time face liveness detection system using YOLOv8 to address these vulnerabilities. Under controlled laboratory conditions, the system achieved exceptional performance metrics: accuracy of 1.0, precision of 1.0, and recall of 1.0, with a mean Average Precision (mAP) of 0.96. However, this study reveals critical insights about the challenges of real-world deployment, including significant performance degradation under poor lighting conditions where genuine faces were misclassified as spoofed images. Compared to existing methods such as Attention-Based Two-Stream CNN (accuracy: 0.91) and Deep Spatial Gradient approaches (accuracy: 0.90-0.92), our system demonstrates superior performance in controlled environments but highlights the persistent challenge of environmental variability in practical applications. These findings emphasize the need for robust preprocessing techniques and diverse training datasets to bridge the gap between laboratory performance and real-world reliability. The study contributes to understanding the limitations of current face anti-spoofing technologies and provides a foundation for developing more robust systems suitable for practical deployment.