Isharah-Selfie Dataset for One-Handed Arabic Sign Language Recognition
October 27, 2026
Isharah-Selfie provides 8,997 video clips of one-handed Arabic Sign Language (ArSL) captured via smartphone front-facing cameras. The dataset covers 1,493 unique sentences across healthcare, education, and transportation domains in unconstrained, mobile-first environments.
HOW THIS AFFECTS YOU
●
builderYou can train more robust sign language models for mobile-first applications and real-world lighting conditions.
●
designerYou can design better UX for accessibility tools that rely on single-handed input.