Abdrah

SpatialRGPT-Bench

A spatial reasoning visual question answering benchmark for evaluating 3D spatial cognition in vision-language models, incorporating RGB and depth images from warehouse, indoor, and synthetic environments with annotated 3D bounding boxes.

Downloads76

Technical Profile

Modalities
rgbdepthlanguage
Environment
warehouseindoorsimulation
Task Types
visual-question-answeringspatial-reasoning
Data Format
HDF5
Part of the SpatialRGPT-Bench family

Access

Need custom rgb data?

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

Request a Sample Pack

Related Datasets