Overview
ShapeNet is a collaborative research project by Princeton, Stanford, and TTIC that provides a comprehensive dataset of 3D shapes. It offers researchers access to annotated 3D models for use in various fields like computer graphics, computer vision, and robotics. The dataset includes ShapeNetCore, featuring single, clean 3D models with verified category and alignment annotations covering 55 common object categories. ShapeNetSem, a smaller but densely annotated subset, offers models with real-world dimensions, material composition estimates, and volume/weight estimates. ShapeNet aims to accelerate research by providing a standardized, large-scale resource for training and evaluating 3D shape analysis algorithms and models. Access to ShapeNet requires user registration, and the data is available for download.
Common tasks