First-person and household activity data for robotics, embodied AI and multimodal models.
First-person and household activity data for robotics, embodied AI and multimodal models.
Needed first-person video and household activity data with spoken narration to train robotics and embodied AI models.
Egocentric video, human-object interaction, household tasks, narration and spoken descriptions, task-based collection and custom annotation.
Task protocol design, participant recruitment and consent, first-person video collection, narration capture and multi-level QA.
First-person / egocentric video, household activities, human-object interaction, narration and spoken descriptions.
Captures first-person video and narration per task protocol.
Validates task segmentation, interaction labels and narration accuracy.
Final validation against acceptance thresholds before delivery.
Describe your task set, environment and annotation requirements.