3D Shape Modeling and Analysis

Computational approaches to 3D shape modeling aim to represent, reconstruct, and analyze the geometry of physical objects from data such as point clouds, meshes, and images. The ability to accurately recover shape from sparse or incomplete observations underpins applications ranging from robotics and medical imaging to autonomous vehicles and digital fabrication. Deep learning has recently shifted the field by allowing models to learn rich shape representations directly from data, enabling tasks like segmentation, classification, and single-image reconstruction that were previously intractable at scale. Active open questions include how to achieve robust generalization across shape categories with limited labeled data, and how to efficiently parameterize and deform continuous surfaces in ways that remain geometrically faithful under large transformations.

Works

49,772

Total citations

629,364

Keywords

Deep LearningPoint Clouds3D ReconstructionMesh SegmentationShape RepresentationNeural Radiance Fields