Empowering Deaf Community in Healthcare Communication: 1D-CNN-Based Algerian Sign Language Recognition System

Language is the fundamental means of communication in our daily lives, encompassing various aspects of individual behavior. However, for the deaf community, the absence of participation in the language communication system due to hearing impairment results in a significant disability that leads to isolation. This communication barrier has detrimental effects, particularly in healthcare, where communication between patients and doctors is hindered. In this paper, to support the public health of deaf and hard-of-hearing individuals in Algeria, we created an Algerian medical word recognition system based on one-dimensional Convolutional Neural Networks (1D-CNN) trained on a collected dataset of 100 videos. Our model utilizes body landmark points extracted from video frames, enhanced with novel features capturing the spatial relationships and angles between these points. Notably, our proposed 1D-CNN model achieved an exceptional accuracy of 100% for Algerian word recognition. This result underscores the effectiveness of our proposed model as a powerful tool in breaking down communication barriers for the deaf community in Arabic-speaking countries, with a particular focus on Algeria.