Differentiable Optics via Ray Tracing

gradoptics is a ray tracing based optical simulator built using PyTorch [1] to enable automatic differentiation.

The API is designed similar to rendering softwares, and has been heavily inspired by Physically Based Rendering (Pharr, Jakob, Humphreys) [2].

Installation

pip install gradoptics

Work in progress

• Currently, some optical element normals are aligned with the optical axis -> more general orientations in progress

• Currently, monochromatic -> no chromatic aberrations

Project History

This project was started in 2020 by Michael Kagan and Maxime Vandegar at SLAC National Accelerator Laboratory.

Feedback and Contributions

Please use issues on GitHub for reporting bugs and suggesting features (including better documentation).

We appreciate all contributions. In general, we recommend using pull requests to make changes to gradoptics.

Support

gradoptics was developed in the context of the MAGIS-100 experiment

References

[1] A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, et al. PyTorch: An imperative style, high-performance deep learning library. In NeurIPS, 2019.

[2] Matt Pharr, Wenzel Jakob, and Greg Humphreys. 2016. Physically Based Rendering: From Theory to Implementation (3rd ed.). Morgan Kaufmann Publishers Inc.