🚧 African Actions Annotated Videos for AI (Health Domain) 🚧

A multimodal, egocentric video dataset of everyday activities performed in Nigeria, built to bring African contexts into the data used to train embodied, agentic, and multimodal AI across many domains. The "Health domains" subset contains 10 egocentric videos and 1,666 time-stamped action-object-tool segments in the "health" domain. Every activity is broken into short segments. Each segment labels the action being performed, the object it is performed on, the tool used, and a natural-language description in English and in an indigenous Nigerian language (Igbo, Hausa, or Yoruba). In total the dataset provides 1,666 annotated segments spanning 652 distinct action–object units, making the "how" of African tasks in the health domain — not just the "what" — available to machines.