Comparison Research: YOLO (You Only Look Once) Model for Indonesian Sign Language Detection Reducing Communication Inequalities
Brandon Steven, Calista Vannia Nathalie, Christoper Gerry Johanes, Puti Andam Suri · 2024
This research explores the effectiveness of two YOLO (You Only Look Once) models, which are YOLOv8 and YOLOv9 in detecting Indonesian Sign Language gestures. The research focuses on AI (Artificial Intelligence) systems to bridge communication and prevent isolation among the DHH (Deaf and Hard of Hearing) community. The YOLOv8n and YOLOv9c from Ultralytics were compared in terms of their performance from preprocessing the dataset into the model compatibility followed by evaluating them by using the standard AI metrics. This research is significant as it contributes to developing more solutions for the tenth SDGs (Sustainable Development Goals) that is to promote equality. The result of this research clearly indicates that YOLOv9 is much better than YOLOv8 as a model in case of detecting Indonesian Sign Language gestures. This research not only can greatly enhance AI accuracy and dependability, but also opening doors for future improvements in computer vision and social equality.