Learning Sign Language Using Single Shot Detector (SSD) And Mobilenet

• Main Author: Nik Ahmad Farihin, Mohd Zulkifli
• Format: Undergraduates Project Papers
• Language: English
• Published: 2023
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://umpir.ump.edu.my/id/eprint/40906/1/CB20179.pdf; http://umpir.ump.edu.my/id/eprint/40906/

Sign languages are a form of communication used by the deaf and hard-of-hearing community. Malay Sign Language (MSL) is the official sign language that is practiced in Malaysia to communicate using hand signs and facial expressions. Every sign and its combination have a different meaning, this makes it quite hard for people to just casually pick up Malay Sign Language to learn. Therefore, this study presents an object detection model using Single Shot Detector (SSD) and Mobilenet to detect Sign Language in real time. This model is only trained to detect static signs which didn’t require any complex combination. The dataset consists of 2000 sign images that were collected from a website called Kaggle and collected using a personal camera. For the training, validation, and testing phases, the dataset was divided into 8:1:1 respectively. In conclusion, this thesis has succeeded in developing a real-time and accurate system for MSL recognition using the SSD-Mobilenet model, which can contribute to the field of sign language recognition and help to improve communication access for deaf and hard-of-hearing individuals.