Infinigen is an open-source procedural generator for creating diverse and high-quality 3D scenes. Developed by Princeton Vision & Learning Lab, it is optimized for computer vision research and generates realistic training data. Infinigen uses randomized mathematical rules to generate unlimited variations of shapes, materials, and details. It offers a wide range of generators for natural objects and scenes, with a focus on accurate geometry rather than faking details. Infinigen also provides automatic annotations for various computer vision tasks and encourages community contribution and expansion of its capabilities.
INFINGEN
1 min read