Path Tracing Denoising Based on SURE Adaptive Sampling and Neural Network

Xing, Qiwei; Chen, Chun-Yi

doi:10.1109/access.2020.2999891

Cited by 6 publications

(3 citation statements)

References 45 publications

(53 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SURE has appeared multiple times in computer graphics literature, being used to guide adaptive sampling and to optimize parameters of denoising functions [LWC12,RMZ13,CFS ∗ 18,XC20].…”

Section: Related Workmentioning

confidence: 99%

Progressive Denoising of Monte Carlo Rendered Images

Firmino

Frisvad

Jensen

2022

Computer Graphics Forum

View full text Add to dashboard Cite

a) input 512spp (b) input crop (c) denoised (d) ours (e) referenceFigure 1: Loss of detail when denoising, (b) to (c), and subsequent recovery with our approach (d). By incorporating error estimates of (b)and (c), our method is able to infer a per-pixel mixing parameter used to interpolate between the two images and reduce the overall error.Here we used Intel® Open Image Denoise (OIDN) to denoise the image, albedo and normal auxillary features included.

show abstract

“…SURE has appeared multiple times in computer graphics literature, being used to guide adaptive sampling and to optimize parameters of denoising functions [LWC12,RMZ13,CFS ∗ 18,XC20].…”

Section: Related Workmentioning

confidence: 99%

Progressive Denoising of Monte Carlo Rendered Images

Firmino

Frisvad

Jensen

2022

Computer Graphics Forum

View full text Add to dashboard Cite

show abstract

“…Xing and Chen [14] also adapted a parameter estimation network to address noise artifacts from path tracing. The method contains sampling and reconstruction stages.…”

Section: Parameter Predictionmentioning

confidence: 99%

A survey on deep learning-based Monte Carlo denoising

Huo

Yoon

2021

Comp. Visual Media

View full text Add to dashboard Cite

Monte Carlo (MC) integration is used ubiquitously in realistic image synthesis because of its flexibility and generality. However, the integration has to balance estimator bias and variance, which causes visually distracting noise with low sample counts. Existing solutions fall into two categories, in-process sampling schemes and post-processing reconstruction schemes. This report summarizes recent trends in the post-processing reconstruction scheme. Recent years have seen increasing attention and significant progress in denoising MC rendering with deep learning, by training neural networks to reconstruct denoised rendering results from sparse MC samples. Many of these techniques show promising results in real-world applications, and this report aims to provide an assessment of these approaches for practitioners and researchers.

show abstract

“…Recently, machine learning-based denoising approaches [6,[13][14][15] have been demonstrated to provide more effective means to denoise the images. However, according the Nyquist sampling theorem, these methods are limited by the minimum number of samples per pixel [16] . That is to say, if the samples number is less than the minimum value, they could not obtain the denoising images with better images.…”

Section: Introductionmentioning

confidence: 99%

Accelerated Path Tracing With GAN and Matrix Completion

Xing

Chen

2021

IEEE Access

Self Cite

View full text Add to dashboard Cite

Denoising-based techniques have recently been shown to be effective for accelerating path tracing rendering methods. However, there remains a problem which is input images need the minimum necessary samples number in order to ensure the quality of the output. In this paper, we propose a new accelerated path tracing approach with generative adversarial networks(GAN) and matrix completion. Unlike the methods based on denoising with neural network, we randomly render part of pixels of input image, which are much less than other methods. Next, we utilize the trained GAN to pre-complete the initializing missing pixels. Because of the accuracy and fast-convergence of GAN, our pre-completion results are more accurate than other methods. Then, according to the results of pre-completion, we present the pre-completed images as a low-rank matrix and make use of the matrix completion to recovers missing values accurately even in high details. To improve the efficiency of solving matrix completion, we modified the original weighted nuclear norm minimization with a parameter adjustment(PAWNNM) strategy. The result shows better visual quality, texture details and convergence efficiency than the state-of-the-art acceleration methods, especially the methods based on denoising with neural network.

show abstract

Path Tracing Denoising Based on SURE Adaptive Sampling and Neural Network

Cited by 6 publications

References 45 publications

Progressive Denoising of Monte Carlo Rendered Images

Progressive Denoising of Monte Carlo Rendered Images

A survey on deep learning-based Monte Carlo denoising

Accelerated Path Tracing With GAN and Matrix Completion

Contact Info

Product

Resources

About