Akaispacelabs2026apache-2.0

Akai_Ego_100_Hrs_v1

A collection of approximately 100 hours of egocentric recordings captured across diverse real-world activities and environments, with synchronized RGB video and inertial measurements for embodied AI and multimodal learning research.

Downloads32
Hours100
Likes1

Why This Matters for Physical AI

This egocentric dataset with synchronized video and IMU data enables training of embodied foundation models and vision-language-action systems that can learn human movement and interaction patterns from real-world experiences.

Technical Profile

Modalities
rgbimu
Environment
indoor
Total Hours
100
Data Format
mp4, csv, json, parquet
Annotation Types
metadata
License
apache-2.0
Part of the Akai_Ego_100_Hrs_v1 family

Access

Need custom rgb data?

Claru builds purpose-built datasets for indoor applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets