felixfabricius2026cc-by-4.0

Robot chess commentator: square dataset

Square crops of a real chess board photographed by a Reachy Mini robot, labeled with chess pieces on each square. Contains 23,744 labeled squares from 371 board positions for board-state recognition tasks.

Downloads0
Episodes371

Why This Matters for Physical AI

This dataset enables training vision models for robotic perception of structured environments, specifically board-state recognition which is foundational for robots performing multi-step manipulation tasks in constrained domains.

Technical Profile

Modalities
rgbsegmentation_mask
Robot Embodiments
Reachy Mini
Environment
lab
Task Types
board-state-recognitionimage-classificationmove-estimation
Episodes
371
Data Format
parquet
Annotation Types
action_labelssegmentation
License
cc-by-4.0
Part of the Robot chess commentator family

Access

Need custom rgb data?

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

Request a Sample Pack

Related Datasets