AutDB - Privacy‑First Benchmark for Video‑Based ASD Screening
Abstract
A reproducible, privacy‑preserving baseline that harmonizes two datasets - MMASD (video‑derived skeleton/flow) and Engagnition (E4 wristband + annotations) - into a unified table and evaluates transportability (LODO) and IID performance on a shared proxy target.
Fields of Study & Subjects
Privacy-Preserving Technologies in Data
Internet Traffic Analysis and Secure E-voting
Face recognition and analysis
Publication details
- Journal / source
- OSF Preprints (OSF Preprints)
- Publication date
- 7 January 2026
- Document type
- Preprint