Video · Multimodal / AI & Machine Learning
Egocentric Video Dataset | 100,000 hours of First-Person Videos for Physical AI, Robotics & Egocentric Vision
Advertised by providerLicensing rights not independently verifiedPrice on request
Overview
Egocentric Video Dataset | 100,000 hours of First-Person Videos for Physical AI, Robotics & Egocentric Vision is a multimodal dataset. The provider states a volume of 100,000 hours. Typical applications include robotics & embodied AI. Relevant sectors: AI & Machine Learning, Education and Manufacturing & Industrial. Pricing is not published and is provided on request.
Primary applications
What the listing offers
- Stated volume of 100k hours.
The Egocentric Residential Task Video Dataset is a 100,000-hour collection of high-quality first-person videos recorded in real residential environments across Brazil at 1080p resolution or higher. The dataset captures natural household chores from the human point of view, preserving hand-object interactions, room context, object placement, surfaces, tools, and multi-step task flows. The dataset is built for Physical AI, robotics, egocentric vision, household automation, manipulation learning, and embodied AI. It focuses on non-cooking residential tasks including cleaning, laundry, organizing, bathroom maintenance, dish handling, pet-area upkeep, plant care, storage organization, basic maintenance, and household navigation. Each recording includes egocentric video paired with structured annotations for task type, action segments, room type, household objects, tools, surfaces, hand-object interactions, pickup events, placement events, open and close events, wiping and scrubbing events, task phases, completion states, and before-and-after household state changes. The annotation structure helps models learn not only the visible action, but also the role of that action inside a longer residential workflow. The dataset includes 80 non-cooking household task categories across bedrooms, bathrooms, kitchens, laundry areas, living rooms, closets, storage spaces, balconies, patios, and utility areas. Coverage includes making beds, changing bedsheets, folding laundry, loading and unload
Technical specifications
Related datasets
More in Video & Physical AI →Video · MultimodalAdvertised
Industrial First-Person Video Dataset | 50,000 Hours of Warehouse & Workplace Tasks - Egocentric Data
The Industrial First-Person Video Dataset is a 50,000-hour collection of egocentric workplace video capturing real industrial, warehouse, logistics, retail,…
- Language
- —
- Volume
- 50k h
- Domain
- AI & Machine Learning +2
- Use case
- Alignment & RLHF
Price on request