Lightweight American Sign Language and Gesture Recognition using YOLOv8

Hand gestures are a kind of nonverbal communication due to their inherent presence in interpersonal conversations, serving as a natural means for individuals to connect with one another. Hand gesture recognition pertains to the cognitive process of understanding and categorizing significant motions generated by an individual's hands. This methodology is used to engage in interpersonal communication with individuals. Hand Hand sign recognition is a challenging task due to the variability of hand gestures, background clutter, and occlusion. Using DeepSORT and YOLOv8, we suggest a simple method for hand sign recognition in this paper. A tracking algorithm called DeepSORT uses deep learning to track multiple objects in a video even when they are obscured. To find hands in a video, use the quick and precise object detection algorithm YOLOv8. First, YOLOv8 is used by our system to find hands in videos. The hands are then monitored over time using DeepSORT. The tracked hands are subsequently divided into various hand signals. Real-time hand sign recognition is possible with our system, which is a lightweight solution. It is also sturdy.