“…The choice of Arabic script is motivated by its unique characteristics and the need for efficient recognition solutions in regions where Arabic is widely used [33]. The model employs a Convolutional Neural Network (CNN) [27], [61] architecture, optimized for TinyML environments, to accurately decode 2D gesture inputs. This approach not only tackles the challenges posed by Arabic script but also leverages the advantages of TinyML to create an effective and efficient solution.…”