# Infinite Photorealistic Worlds using Procedural Generation

**Authors:** Alexander Raistrick, Lahav Lipson, Zeyu Ma, Lingjie Mei, Mingzhe Wang, Yiming Zuo, Karhan Kayan, Hongyu Wen, Beining Han, Yihan Wang, Alejandro Newell, Hei Law, Ankit Goyal, Kaiyu Yang, Jia Deng  
**arXiv:** [2306.09310](https://arxiv.org/abs/2306.09310)  
**Date:** 2023-06-15  
**Summary page:** https://www.summarizepaper.com/en/paper/2306.09310v1/

## Key points

- Infinigen is a procedural generator of photorealistic 3D scenes of the natural world.
- It was introduced by a team of researchers including Alexander Raistrick, Lahav Lipson, and Jia Deng.
- Infinigen generates every asset from shape to texture via randomized mathematical rules without any external source.
- The tool covers objects and scenes in the natural world such as plants, animals, terrains and natural phenomena like fire, cloud, rain and snow.
- Infinigen can generate unlimited diverse training data for a range of computer vision tasks such as object detection, semantic segmentation, optical flow and 3D reconstruction.
- The authors expect Infinigen to be useful not only for computer vision research but also beyond.
- It has been accepted at CVPR 2023 with the camera-ready version available on arXiv.org along with videos, code and pre-generated data.

## Summary

Infinigen is a revolutionary procedural generator of photorealistic 3D scenes of the natural world. It was introduced by a team of researchers including Alexander Raistrick, Lahav Lipson, and Jia Deng. Infinigen is entirely procedural; every asset from shape to texture is generated from scratch via randomized mathematical rules without any external source. This allows for infinite variation and composition. The tool covers objects and scenes in the natural world such as plants, animals, terrains and natural phenomena like fire, cloud, rain and snow. It can be used to generate unlimited diverse training data for a range of computer vision tasks such as object detection, semantic segmentation, optical flow and 3D reconstruction. The authors expect Infinigen to be useful not only for computer vision research but also beyond; it has been accepted at CVPR 2023 with the camera-ready version available on arXiv.org along with videos, code and pre-generated data.