Yangyihuicc-by-nc-sa-4.0

Sequence Object Navigation R5

A non-commercial research dataset of 13,647 egocentric RGB videos with 18,249 object-goal navigation windows derived from SpatialVID-HQ, designed for video pretraining and visual navigation representation learning.

Downloads0
Episodes13647

Why This Matters for Physical AI

This dataset provides large-scale egocentric video data with spatial instruction grounding for training visual navigation models and embodied AI systems to understand spatial relationships and follow natural language navigation directives.

Technical Profile

Modalities
rgb
Robot Embodiments
quadruped
Environment
indoor
Task Types
navigationobject_goal_navigation
Episodes
13647
Data Format
tar
Annotation Types
language_instructionsbounding_boxessegmentation
License
cc-by-nc-sa-4.0
Part of the SpatialVID-HQ 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