基于深度学习的光线追踪降噪方案

作者

• 方天宬 [email protected]
• 陆徐东 [email protected]

简介

本项目是SJTU21-22学年CS403 计算机图形学课程的大作业，本文是我们小组项目的参考文档。我们主要实现了以下功能：

1. 在实时降噪方面，我们复现并改进了SIGGRAPH17的RDAE
2. 在离线降噪方面，我们复现了SIGGRAPH19的文章SBMC

在两种方法上，我们都取得了良好的效果。

文件夹

report      报告
siggraph17  RDAE改进方案
siggraph19  SBMC论文复现

更详细的指南请见对应文件夹下README

Ray tracing denoising methods based on deep learning

Author

• Tiancheng Fang [email protected]
• Xudong Lu [email protected]

Introduction

This project is a major assignment for the CS403 Computer Graphics course of SJTU 21-22 academic year. This tutorial is a reference document of our group project. We mainly realize the following functions:

• In terms of real-time denosing, we reproduce and improve the rdae of siggraph17
• In the aspect of offline denosing, we reproduce the article SBMC of siggraph19

We have achieved good results in both methods.

Directory

report      # report
siggraph17  # RDAE
siggraph19  # SBMC

For more detailed guidance, please refer to README under the corresponding folder.