Indoor Egocentric Chores Video Dataset

I need a set of high-quality egocentric videos recorded indoors while you perform everyday household chores. The footage will be used to train computer-vision models, so natural head- or chest-mounted point-of-view is essential and camera shake must stay true to a normal human gait.

Scope of recording
• Location: any typical indoor home setting—kitchens, living rooms, bedrooms, laundry rooms, garages, etc.
• Activities: cleaning, cooking, washing dishes, folding clothes, making beds, organizing shelves, sweeping, vacuuming and similar routine chores. Feel free to vary lighting conditions and rooms so the final dataset is diverse.
• Perspective: first-person only; no third-person shots.
• Duration: target 30–60 minutes of usable footage per participant, delivered as multiple shorter clips is fine.
• Format: 1080p or higher, 30 fps+, H.264 or H.265, with original audio. No post-processing besides trimming obvious dead time.
• Privacy: record only yourself and items you own; remove any personal identifiers you don’t wish to share. Written consent that I can use the footage for machine-learning research is required.

Deliverables
1. Raw video files.
2. A simple CSV/JSON manifest listing file name, activity label, room type, and approximate timestamp ranges for each chore segment.
3. Signed consent form granting research and commercial ML usage rights.

Acceptance criteria
• Footage is stable, correctly exposed, and fully from the wearer’s perspective.
• Every clip is clearly labelled in the manifest.
• Total runtime meets or exceeds 30 minutes of varied chores.

If this first round goes smoothly, I may commission additional indoor scenarios or longer recordings.