Occluded 3D Point Clouds of Stanford Bunny, Teapot, and Cow

Three 3D point cloud models, each provided in both occluded and unoccluded versions, were synthesized from polygonal models. The dataset includes the Stanford bunny, Utah Teapot, and OSG Cow, with shapes obscured by amorphous point clouds to study visual perception. It was created by the Department of Energy and last updated in March 2026.

Use Cases

• Train 3D shape recognition models to distinguish the Stanford bunny, Utah Teapot, or OSG Cow from within occluding point clouds.
• Benchmark visual perception algorithms by comparing performance on occluded versus unoccluded versions of the same 3D model.
• Study the effectiveness of 3D scatterplot rotation versus 2D axis-aligned projections for human identification of occluded shapes.
• Develop point cloud processing techniques for feature extraction from organically structured, amorphous point clouds.

Strengths

• Includes three distinct, well-known 3D models (Stanford bunny, Utah Teapot, OSG Cow) for comparative analysis.
• Each model is provided in both occluded and unoccluded versions, enabling controlled perceptual studies.
• Dataset provenance is from the Department of Energy and was updated in March 2026.

Limitations

• Very small scale, containing only three distinct object models, limiting statistical analysis.
• No columnar metadata or sample data is provided, complicating automated parsing and analysis.
• Data format is BIN files without documented structure, requiring specialized tools for interpretation.

Provenance

Source

Department of Energy

Collection Method

Synthesized from polygonal models via point-sampling and occlusion with amorphous point clouds.

Freshness

Last updated March 2026.