TransBiolab: A Real-World Multi-View Dataset of Cluttered Transparent Biomedical Objects

Researchers released TransBiolab, a real-world calibrated multi-view RGB-D dataset for transparent biomedical objects in cluttered lab scenes.

It includes 161,315 frames, 98 scenes, and over 1.03 million instance annotations across 15 laboratory object types.

The dataset comes with benchmarks for segmentation, depth completion, and 6D pose estimation.

TransBiolab addresses a major data gap for autonomous biomedical labs and robot manipulation with transparent plasticware.