Arabic Sign Language Recognition Chatbot for Decision Making in Healthcare

This paper presents the development of a healthcareoriented chatbot system designed to facilitate communication for hearing-impaired individuals by recognizing Arabic sign language (ArSL). To identify the most effective deep learning model for this application, we conducted a comparative analysis of four widely used pretrained image classifiers-VGG16, ResNet50, InceptionV3, and EfficientNetB0-on the RGB Arabic Alphabet Sign Language Dataset (AASL). Through extensive experimentation, we selected the best model based on accuracy for integration into the chatbot system, which will process sequences of ArSL gesture images, translate them into sentences, and enable real-time, accessible interactions in healthcare settings. This approach underscores the potential of deep learning to enhance healthcare accessibility, providing a practical, AI-driven communication solution for hearing-impaired individuals.