1Northeastern University
2The Hong Kong University of Science and Technology (Guangzhou)
3Sichuan University
4Beijing Innovation Center of Humanoid Robotics Co., Ltd.
We present GravLensX, an innovative method for rendering black holes with gravitational lensing effects using neural networks. The methodology involves training neural networks to fit the spacetime around black holes and then employing these trained models to generate the path of light rays affected by gravitational lensing. This enables efficient and scalable simulations of black holes with optically thin accretion disks, significantly decreasing the time required for rendering compared to traditional methods. We validate our approach through extensive rendering of multiple black hole systems with superposed Kerr metric, demonstrating its capability to produce accurate visualizations with significantly 15× reduced computational time. Our findings suggest that neural networks offer a promising alternative for rendering complex astrophysical phenomena, potentially paving a new path to astronomical visualization.
We train separate neural networks to model the null-geodesic geometry in each region—covering multiple near fields and a far field. Then we perform ray tracing within these regions to reconstruct the complete light paths.
The trained networks fit the ground truth null-geodesic paths very well for both near and far fields.
Accretion Disk Rendering
Sky Sphere Rendering
Full Rendering & A Comparison with Flat Space
Our method achieves high-fidelity black-hole rendering with gravitational lensing effects, proving neural networks' potential to accelerate complex astrophysical visualizations.
If you find our work useful, please consider citing it as:
@inproceedings{sun2025learning,
title={Learning Null Geodesics for Gravitational Lensing Rendering in General Relativity},
author={Sun, Mingyuan and Fang, Zheng and Wang, Jiaxu and Zhang, Kunyi and Zhang, Qiang and Xu, Renjing},
booktitle={2025 IEEE/CVF International Conference on Computer Vision (ICCV)},
year={2025},
organization={IEEE}
}