Classroom Distraction Evaluation Dataset (CDED-7) — v1.1.0-dataset: seven self-collected classroom videos, per-video annotations, and a privacy-aware derived annotation package

This dataset is the version-of-record for the self-curated evaluation set used in the manuscript "Interpretable classroom distraction estimation via tracking-to-analysis continuity" (PeerJ Computer Science). It supports the tracking-to-analysis continuity framework described in the paper and is released for research and non-commercial use only. What's in the archive --------------------- * videos/class_1.mp4 ... class_7.mp4 Seven self-collected classroom videos recorded at the School of Artificial Intelligence, Luoyang Normal University, China. Each video is 1920x1080, 25-30 FPS, shot with fixed surveillance-grade IP cameras mounted at instructor- station height. No audio channel is included. See Section 6.5 Dataset Construction Protocol of the README for the full recording protocol, inclusion / exclusion criteria, and anonymisation safeguards: https://github.com/fangsheng223/classroom-distraction-tracker * derived_release/ A privacy-aware reproducibility package: per-frame bounding boxes, stable person IDs, and a binary attention label (Focused / Distracted) for all seven videos. Raw frames, audio, face images, and original local file paths are intentionally omitted. * submission_assets/ Manuscript figures and supplementary LaTeX tables supporting the peer-reviewed work. Licence ------- Research and non-commercial use ONLY. The full terms, including the list of permitted uses (academic research, method development, reproducibility), prohibited uses (commercial use, re-identification, face recognition, biometric identification, emotion scoring, surveillance, profiling, punitive decision-making), the consent statement, the downstream privacy obligations, and the withdrawal procedure are specified in DATASET_LICENSE.md: https://github.com/fangsheng223/classroom-distraction-tracker/blob/main/DATASET_LICENSE.md Where this license and any other applicable terms disagree, the stricter restriction applies. Citation -------- Please cite BOTH the PeerJ Computer Science manuscript AND this Zenodo record (the DOI at the top of this page) when using this dataset. The corresponding source code, derived annotations, and the tagged GitHub Release are mirrored at: https://github.com/fangsheng223/classroom-distraction-tracker/releases/tag/v1.1.0-dataset