International Research Journal of Education and Technology

Improving communication for the deaf and hard-of- hearing is the aim of machine learning-based sign language translation (SLT). This study looks at the most recent advancements in deep learning techniques, such as Transformers, Recurrent Neural Networks (RNNs), and Convolutional Neural Networks (CNNs), for translating sign language gestures into text or voice. Significant challenges include the range of sign languages, real-time processing, and dataset limitations. Despite these difficulties, translation accuracy has significantly increased due to developments in computer vision and natural language processing. Future research could focus on developing real-world, accessible SLT systems, enhancing the interpretability of models, and integrating multimodal sensors.

Key Words: — Sign Language Translation, Machine Learning, Deep Learning, Convolutional Neural Networks, Recurrent Neural Networks, Transformers, Gesture Recognition, Computer Vision, Accessibility, Natural Language Processing, Multi-modal Sensors.