chenbr72026cc-by-nc-4.0

TransFrag27K: Transparent Fragments Dataset

A large-scale synthetic dataset of 27,000 images and masks for transparent glass fragment segmentation, generated via automated Blender pipeline with diverse lighting and backgrounds. Designed for autonomous reassembly tasks using visual-tactile fusion.

Downloads63
Episodes27000
Likes1

Why This Matters for Physical AI

This dataset enables development of perception systems for transparent object manipulation and reassembly, addressing a critical challenge in robotic handling of visually difficult materials with refractive properties.

Technical Profile

Modalities
rgbimage
Environment
simulation
Task Types
segmentationreassembly
Episodes
27000
Annotation Types
segmentationbinary_masks
License
cc-by-nc-4.0
Part of the TransFrag27K: Transparent Fragments Dataset family

Access

Need custom rgb data?

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

Request a Sample Pack

Related Datasets