This study aims to develop a machine learning-based application capable of detecting hand gestures and patterns in Indonesian Sign Language (BISINDO). Sign language plays a crucial role in non-verbal communication, particularly for individuals with speech impairments like the deaf. However, the challenge of comprehending sign language often inhibits interactions between the deaf and others. In an effort to address this barrier, the research leverages machine learning techniques with a focus on the Convolutional Neural Network (CNN) method, utilizing a dataset annotated with hand gesture landmarks. Landmark information providing detailed positions and shapes of key points on the hand, the CNN model can learn specific features essential for classification. The resulting application aims to bridge communication between the deaf and other individuals who may not understand sign language. By harnessing this technology, a significant improvement in the accuracy of hand gesture classification in sign language is anticipated, thereby strengthening the communication and interaction capabilities of the deaf within their environment.
Copyrights © 2023