SignSpeak - Sign Language Interconverter Using CNN-based Approach

This research paper is oriented towards the creation of a mobile app called SignSpeak, which serves as an all-encompassing solution for converting sign language to text and vice versa. It accomplishes this through a Convolutional Neural Network (CNN) approach. The paper begins with the evaluation of existing tools for sign language interpretation, identification of their limitations, and the elucidation of how SignSpeak offers an effective remedy for these issues. It then provides an extensive account of the system's implementation, covering the app's development, data gathering, model selection, CNN model training, and the integration of the app's front-end with the trained model's back-end. Subsequently, the paper offers an intricate description of the system's functional and operational aspects. Finally, it delves into the achieved results and outlines future research opportunities.