Neural Parametric Mixtures for Path Guiding

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Dong, Honghao, Wang, Guoping, Li, Sheng
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916685068369920
author Dong, Honghao
Wang, Guoping
Li, Sheng
author_facet Dong, Honghao
Wang, Guoping
Li, Sheng
contents Previous path guiding techniques typically rely on spatial subdivision structures to approximate directional target distributions, which may cause failure to capture spatio-directional correlations and introduce parallax issue. In this paper, we present Neural Parametric Mixtures (NPM), a neural formulation to encode target distributions for path guiding algorithms. We propose to use a continuous and compact neural implicit representation for encoding parametric models while decoding them via lightweight neural networks. We then derive a gradient-based optimization strategy to directly train the parameters of NPM with noisy Monte Carlo radiance estimates. Our approach efficiently models the target distribution (incident radiance or the product integrand) for path guiding, and outperforms previous guiding methods by capturing the spatio-directional correlations more accurately. Moreover, our approach is more training efficient and is practical for parallelization on modern GPUs.
format Preprint
id arxiv_https___arxiv_org_abs_2504_04315
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Neural Parametric Mixtures for Path Guiding
Dong, Honghao
Wang, Guoping
Li, Sheng
Graphics
Previous path guiding techniques typically rely on spatial subdivision structures to approximate directional target distributions, which may cause failure to capture spatio-directional correlations and introduce parallax issue. In this paper, we present Neural Parametric Mixtures (NPM), a neural formulation to encode target distributions for path guiding algorithms. We propose to use a continuous and compact neural implicit representation for encoding parametric models while decoding them via lightweight neural networks. We then derive a gradient-based optimization strategy to directly train the parameters of NPM with noisy Monte Carlo radiance estimates. Our approach efficiently models the target distribution (incident radiance or the product integrand) for path guiding, and outperforms previous guiding methods by capturing the spatio-directional correlations more accurately. Moreover, our approach is more training efficient and is practical for parallelization on modern GPUs.
title Neural Parametric Mixtures for Path Guiding
topic Graphics
url https://arxiv.org/abs/2504.04315