Dates: 2023
Principal Investigator: Dr. Hamzah Luqman
Funded by: SDAIA-KFUPM Joint Research Center for Artificial Intelligence

Description: Continuous Sign Language Recognition (CSLR) translates sign language gestures from video into text, but Arabic sign languages, especially Saudi Sign Language (SSL), remain under-researched, leaving the Saudi deaf community facing communication barriers. Developing Saudi CSLR is challenging due to SSL’s complex grammar, simultaneous use of hand, body, and facial expressions, signer variations, and issues like niqab obscuring facial cues. To address this, a project is building SSL video datasets and developing real-time, lightweight CSLR models that work across signers and environments using smartphones for data collection. The system will also use few-shot learning to overcome limited data and include an Arabic-to-SSL translation component with avatars or animations for communication support.