HUMAN MOTION DETECTION UNTUK TERJEMAH ‎‎BAHASA ISYARAT MENGGUNAKAN METODE RANDOM ‎‎FOREST

‎ ‎Communication is something extremely vital for social beings such as ‎humans. However, not all humans can perform it normally, and one of the ‎reasons is due to disabilities. Therefore, Sign Language is required in order to ‎maintain communication. Sign language is a language that prioritizes body ‎language and lip movements, not sound, for communication. Deaf individuals are ‎a group that frequently utilizes this language, usually by combining hand ‎movements, finger movements, body movements, and facial expressions to ‎express their thoughts. However, not everyone understands how to communicate ‎using sign language. Hence, a sign language translator detection system is ‎needed to facilitate communication between those who understand sign language ‎and those who do not. This research employs the Random Forest method with the ‎Holistic architecture detected by MediaPipe, using pose estimation to obtain ‎landmarks/keypoints for each input image on the hands, face, and body. The ‎obtained results create a sign language detection system that can be operated in ‎real-time with a high level of accuracy, averaging 97% in each classification.‎