AkwaabaSign

The AkwaabaSign dataset contains 5,750 high-resolution (1920×1080, 30 fps) videos of 115 Ghanaian Sign Language (GhSL) words, performed by five signers with 10 variations each, organized hierarchically for education and health contexts. Recorded at KNUST with expert validation, it supports sign language recognition and translation for machine learning applications in Ghana’s hearing-impaired community.