本文探讨了深度学习技术如何革新实时渲染与光照计算领域，介绍了一系列使用卷积神经网络、生成对抗网络等进行屏幕空间阴影、动态全局光照、间接光照、超采样及降噪的前沿研究。
Deep shading: convolutional neural networks for screen space shading
https://arxiv.org/pdf/1603.06078.pdf
Deep Illumination: Approximating Dynamic Global Illumination with Generative Adversarial Networks
2020-Lightweight Bilateral Convolutional Neural Networks for Interactive Single-bounce Diffuse Indirect Illumination
Neural-Supersampling-for-Real-time-Rendering
2019-One-Shot Radiance Global Illumination Using Convolutional
Efficient screen space subsurface scattering Siggraph 2018
Neural-Supersampling-for-Real-time-Rendering
Spatiotemporal reservoir resampling for real-time ray tracing with dynamic direct lighting
一些降噪的论文
2016-EG-Nonlinearly Weighted First-order Regression for Denoising Monte Carlo Rednerings
2017-EGSR-Bayesian Collaborative Denoising for Monte Carlo Rendering
2017-SIGGRAPH-KPCN-Kernel-Predicting-Convolutional-Networks-for-Denoising-Monte-Carlo-Renderings
2019-SIGGRAPH-Adversarial Monte Carlo Denoising with Conditioned Auxiliary Feature Modulation
论文资源
扫一扫
561
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
