The first large-scale word-level Sinhala Sign Language dataset, paired with an energy-efficient end-to-end recognition pipeline.
SSL400 is the first large-scale, word-level dataset for Sinhala Sign Language recognition — addressing a major resource gap for an under-served language community. Alongside the dataset, I developed an end-to-end recognition pipeline focused on energy-efficient modelling.
Accessible AI for sign-language users requires both data and efficient models that can run on modest hardware. SSL400 provides the foundation, and the recognition pipeline demonstrates practical, energy-aware deployment.