Malaysian sign language detection using SSD real-time object detection / Nurfarah Idayu Mohamad Fauzi ... [et al.]

Sign language is a natural, visually focused, nonverbal communication channel that enables communication between individuals using facial or physical emotions, postures, and a set of gestures. It is mostly used to communicate with individuals who are deaf or have a hearing impairment. There are several sign languages around the world, just as there are various spoken languages. Malaysian Sign Language is one of them. However, sign language is not compulsory for us to learn, yet some people do not even recognize sign language. This makes it difficult for individuals with disabilities to communicate with others. Thus, this application is developed to overcome the latter issue. By using this application, users will be assisted to identify the sign language; the application acts as the bridge in closing the gap between the disabled and normal individuals in communicating with each other. Single- shot detector (SSD) Algorithm is implemented to perform the detection. The project used the TensorFlow Lite Model Maker library to simplify the process of training a TensorFlow Lite model using a custom dataset. This project utilizes the Agile framework in project development. The result shows that this project achieved a detection accuracy of 75.2%.