Egocentric Video Data Capture

I’m assembling a real-world, first-person video corpus that will be fed directly into machine-learning pipelines. To make the dataset as useful as possible, I need authentic egocentric footage shot in both indoor and outdoor settings as you go about ordinary daily routines—preparing meals, tidying up, commuting, shopping, working at a desk, walking the dog, and other everyday tasks.

What you’ll do
Record yourself from a head-mounted or chest-mounted camera so the viewpoint is unmistakably first person. Keep the camera rolling naturally; no need for scripted dialogue or staged scenes. Only daily-routine activities are required, so there’s no expectation to film sports, workouts, or large social gatherings.

Key deliverables
• Raw, unedited video files (MP4 or MOV)
• A brief text file for each clip noting date, location type (e.g., kitchen, park), and a one-line activity label
• Confirmation that everyone who appears has given consent

Technical notes
1080p or higher, 30 fps+, natural lighting where possible, and clear audio all help downstream model training. Please avoid watermarks or on-screen graphics.

I release payment for each accepted batch, so you can upload footage incrementally. If anything is unclear—gear choice, privacy steps, or file-transfer options—just let me know and I’ll get back to you quickly.