SMILE Continuous DSGS Corpus L2 Phase 2

This dataset comprises the multi-phase SMILE II learner corpus of DSGS, which includes multi-camera video recordings of L2 learners across four data collection phases. Each session includes calibrated recordings, task-based signing exercises, and pose-estimation outputs derived from the video material. All participants completed learning tasks, which enable longitudinal analyses of manual and non-manual components, language and fluency development, and learner variability. The resource supports research in sign language linguistics, learner corpus studies, sign language technology, and the development of AI-driven learning and assessment tools.

Swiss German Sign Language, Sign Language Learner Corpus, L2 Corpus, Automatic Sign Language Assessment, Sign Language Recognition

The dataset was validated through systematic internal checks during data preparation. All filenames, folder structures, metadata entries, and pose-estimation outputs were cross-checked for completeness and consistency. Withdrawn participant data was removed in accordance with consent requirements.