Project 608: Egocentric POV Video Data Collection
Budget: $8 – $15 USD
We are looking for freelancers from the USA and LATAM regions to participate in a large-scale egocentric (first-person point-of-view) video collection project focused on capturing genuine household and industrial activities. The collected footage will help train and improve AI/ML models related to human activity recognition, hand-object interaction, robotics applications, and real-world scenario understanding. Participants will be compensated at $10 per approved video hour for USA contributors and $7 per approved video hour for LATAM countries contributors.
Freelancers will record themselves performing real-world household and industrial activities, including cleaning, organizing, maintenance, logistics, retail, agriculture, manufacturing, construction, automotive services, and other workplace tasks. Recordings must capture genuine task execution with natural movements, visible hand interactions, and continuous activity, with video clips ranging from 5 to 120 minutes. Staged, scripted, or artificial activities will be rejected.
Recordings must be captured using approved devices only, including iPhone 11 or newer, Google Pixel 6 or newer, or Samsung Galaxy S21 or newer, securely mounted on a headstrap for a first-person perspective. Videos must clearly show the contributor’s hands, objects being handled, workspace, and overall activity with stable camera positioning, proper lighting, and clear image quality to meet AI training standards.
To properly streamline sourcing for this project: please complete this form and the Enterprise Talent Success Team will reach out to you via chat here on Freelancer: https://forms.gle/hz8YG3yQndpZyCBs9
Freelancers will record themselves performing real-world household and industrial activities, including cleaning, organizing, maintenance, logistics, retail, agriculture, manufacturing, construction, automotive services, and other workplace tasks. Recordings must capture genuine task execution with natural movements, visible hand interactions, and continuous activity, with video clips ranging from 5 to 120 minutes. Staged, scripted, or artificial activities will be rejected.
Recordings must be captured using approved devices only, including iPhone 11 or newer, Google Pixel 6 or newer, or Samsung Galaxy S21 or newer, securely mounted on a headstrap for a first-person perspective. Videos must clearly show the contributor’s hands, objects being handled, workspace, and overall activity with stable camera positioning, proper lighting, and clear image quality to meet AI training standards.
To properly streamline sourcing for this project: please complete this form and the Enterprise Talent Success Team will reach out to you via chat here on Freelancer: https://forms.gle/hz8YG3yQndpZyCBs9
Related categories:
Video Services
Machine Learning (ML)
Robotics
Video Production
Video Editing
Data Collection