DEMC: A Deep Dual-Encoder Network for Denoising Monte Carlo Rendering
Xin Yang, Wenbo Hu, Dawei Wang, Lijing Zhao, Baocai Yin, Qiang Zhang, Xiaopeng Wei, Hongbo Fu
Embedded Files
Abstract
Abstract
In this paper, we present DEMC, a deep Dual-Encoder network to remove Monte Carlo noise efficiently while preserving details. Denoising Monte Carlo rendering is different from natural image denoising since inexpensive by-products (feature buffers) can be extracted in the rendering stage. Most of them are noise-free and can provide sufficient details for image reconstruction. However, these feature buffers also contain redundant information, which makes Monte Carlo denoising different from natural image denoising. Hence, the main challenge of this topic is how to extract useful information and reconstruct clean images. To address this problem, we propose a novel network structure, Dual-Encoder network with a feature fusion subnetwork, to fuse feature buffers firstly, then encode the fused feature buffers and a noisy image simultaneously, and finally reconstruct a clean image by a decoder network. Compared with the state-of-the-art methods, our model is more robust on a wide range of scenes, and is able to generate satisfactory results in a significantly faster way.
Dual-Encoder Network
Dual-Encoder Network
A deep Dual-Encoder network, used for denoising Monte Carlo renderings. The feature buffers and noisy image are generated from renderer. The feature buffers are firstly fused by a feature fusion sub-network to get a detail map, and then the detail map and noisy image are encoded by the feature encoder and HDR encoder respectively. Finally, the latent representation is decoded to reconstruct a clean image with skip connection from the Dual-Encoder
Experiment Results
Experiment Results
Dataset is Released
Dataset is Released
DUT-MC168 Dataset
Contact information:
Xin Yang, Dalian University of Technology, xinyang@dlut.edu.cn
Hongbo Fu, City University of Hong Kong, fuplus@gmail.com
Please Cite our paper if you use this dataset in your research:
@inproceedings{Yang2019,
title={DEMC: A Deep Dual-Encoder Network for Denoising Monte Carlo Rendering},
author={Yang, Xin and Wenbo, Hu and Dawei, Wang and Lijing, Zhao and Baocai, Yin and Qiang, Zhang and Xiaopeng, Wei and Hongbo, Fu},
booktitle={CVM},
year={2019}
}
Reference
Reference
1. Yang, Xin, Dawei Wang, Wenbo Hu, Lijing Zhao, Xinglin Piao, Dongsheng Zhou, Qiang Zhang, Baocai Yin, Qiang Cai, and Xiaopeng Wei. "Fast Reconstruction for Monte Carlo Rendering Using Deep Convolutional Networks." IEEE Access (2019): 21177-21187.
2. Bako, S., Vogels, T., McWilliams, B., Meyer, M., Novák, J., Harvill, A., Sen, P., Derose, T. and Rousselle, F., 2017. Kernel-predicting convolutional networks for denoising Monte Carlo renderings. ACM Trans. Graph., 36(4), pp.97-1.
Google Sites
Report abuse