ShapeNet: Large-Scale Annotated 3D CAD Model Repository

ShapeNet is a large-scale repository of 3D CAD models representing objects. It contains models from numerous semantic categories organized under the WordNet taxonomy and provides annotations like consistent rigid alignments, parts, and bilateral symmetry planes. The repository was presented by Anne Lynn S. Chang and collaborators.

Use Cases

• Train 3D object classification models using the semantic category labels from the WordNet taxonomy.
• Develop part segmentation algorithms using the annotated part labels for each 3D model.
• Benchmark rigid alignment algorithms using the provided consistent rigid alignment annotations.
• Research on symmetry detection and modeling using the bilateral symmetry plane annotations.
• Use the categorized 3D models for data augmentation in synthetic training data generation for computer vision tasks.

Strengths

• Large-scale collection of 3D models.
• Provides multiple semantic annotations per model, including alignments, parts, and symmetry planes.
• Organizes models under the established WordNet taxonomy.

Limitations

• Specific row count, file size, and temporal coverage are unknown.
• The description mentions 'a multitude' of categories but does not specify an exact count.
• Original data collection methodology and potential geographic or sourcing biases are not detailed.

Provenance

Source

Anne Lynn S. Chang and collaborators.

Collection Method

Collection of 3D CAD models, annotated and organized.

Time Range

null

Freshness

null

Geography

null