Approximating Reflectance Functions using Neural Networks

Gargan, David; Neelamkavil, Francis

doi:10.1007/978-3-7091-6453-2_3

Cited by 11 publications

(8 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our view, there are several factors that have resulted in this progress. First, the idea of continuously parameterizing a field using an MLP without the need to use more complex neural network architectures has simplified the training of fields and reduced the entry barrier [GN98]. Neural fields provide a fundamentally different view of signal processing that is no longer discrete but rather faithful to the original continuous signal.…”

Section: Discussion and Conclusion 14 Discussionmentioning

confidence: 99%

“…Once the ray-surface intersection point has been retrieved, we can calculate the radiance contribution from the point towards the cam-era. This is done with a bidirectional scattering distribution function (BSDF) [CT82,BS12] which can be differentiable or be parameterized as a neural field [GN98]. To aggregate lighting contribution to the point, the most phyiscally accurate method is to solve Kajiya's rendering equation [Kaj86] with a multi-bounce Monte Carlo algorithm [PJH16].…”

Section: Surface Shading and Lightingmentioning

confidence: 99%

See 1 more Smart Citation

Neural Fields in Visual Computing and Beyond

Xie¹,

Takikawa²,

Saito³

et al. 2021

Preprint

View full text Add to dashboard Cite

Advances in machine learning have increased interest in solving visual computing problems using a class of coordinate-based neural networks that parameterize physical properties of scenes or objects across space and time. These methods, which we call neural fields, have seen widespread success in applications such as 3D shape and image synthesis, animation of human bodies, 3D reconstruction, and pose estimation. Rapid progress has led to numerous papers, but a formulation and review of the problems in visual computing has not yet emerged. We provide context, mathematical grounding, and a review of 251 papers in the literature on neural fields. In Part I, we focus on neural fields techniques by identifying common components of neural field methods, including different generalization, representation, forward map, architecture, and manipulation methods. In Part II, we focus on applications of neural fields to different problems in visual computing, and beyond (e.g., robotics, audio). Our review shows the breadth of topics already covered in visual computing, both historically and in current incarnations, and highlights the improved quality, flexibility, and capability brought by neural field methods. Finally, we present a companion website that contributes a living database that can be continually updated by the community. CCS Concepts

show abstract

Section: Discussion and Conclusion 14 Discussionmentioning

confidence: 99%

Section: Surface Shading and Lightingmentioning

confidence: 99%

Neural Fields in Visual Computing and Beyond

Xie¹,

Takikawa²,

Saito³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Gargan & Neelamkavil [12] showed that using an ANN provides excellent approximation performance for a dense BRDF generated using a gonio-reflectometer. Experimenting with different numbers of layers (which affects the ability of the network to either generalise or overfit), they concluded that a three-tier feedforward backpropagation architecture offers the best performance.…”

Section: Neural Network Architecture For Brdf Generationmentioning

confidence: 99%

BRDF Estimation for Faces from a Sparse Dataset Using a Neural Network

Hansen

Atkinson

2013

Computer Analysis of Images and Patterns

View full text Add to dashboard Cite

“…In the 1998 Eurographics Workshop on Rendering, Gargan and Neelamkavil proposed a non-linear representation of BRDFs based on a neural network model 15 . They used standard backpropagation networks with two or three weight layers, linear basis functions and sigmoid activation functions.…”

Section: Neural Networkmentioning

confidence: 99%

Hierarchical Reconstruction of BRDFs using Locally Supported Functions

Noe¹,

Péroche²

2000

Computer Graphics Forum

View full text Add to dashboard Cite

BRDFs are the backbone of realistic rendering algorithms. Analytical models are sometimes ineffective since they frequently cannot represent very particular material characteristics (retro‐reflection, anisotropy, off‐specularity, …). Consequently one might want to use measured BRDF data directly; this leads to solving the problem of BRDF reconstruction. In this paper, we propose a new solution which uses locally supported functions that lead to a hierarchical approach able to take irregular distributions of sampled data into account. This method is not computationally expensive and guarantees a physically valid reconstruction. We also discuss the quality of the reconstruction, introducing some new error parameters.

show abstract

Approximating Reflectance Functions using Neural Networks

Cited by 11 publications

References 19 publications

Neural Fields in Visual Computing and Beyond

Neural Fields in Visual Computing and Beyond

BRDF Estimation for Faces from a Sparse Dataset Using a Neural Network

Hierarchical Reconstruction of BRDFs using Locally Supported Functions

Contact Info

Product

Resources

About