← All resources

Real-time Neural Six-way Lightmaps

Auto-discovered

Paper Advanced

Wei Li, Hanxiao Sun, Tao Huang, Haoxiang Wang, Tongtong Wang, Zherong Pan, Kui Wu

Visit resource ↗

Description

Real-time Neural Six-way Lightmaps is a ACM Transactions on Graphics paper by Wei Li, Hanxiao Sun, Tao Huang, Haoxiang Wang, Tongtong Wang, Zherong Pan, Kui Wu, and collaborators on neural lighting for dynamic participating media in games. The arXiv abstract summarizes a method tuned toward Unreal Engine-style integration costs.

The approach predicts six-way lightmaps from a camera-view guiding map so dynamic smoke and similar media can respond to changing lights, camera motion, and smoke - obstacle interaction within a few milliseconds per frame. It targets game pipelines that already use directional light probes or lightmaps for opaque geometry and need a volumetric counterpart.

Advanced game rendering programmers and researchers working on real-time volumetric lighting and neural acceleration are the readers. Experience with light probes and game-engine integration constraints helps.

The preprint is linked from arXiv with open PDF access. Neural volumetric lighting papers from major Chinese game-technology research groups appear increasingly in real-time GI conference sessions.

Link last verified accessible on 2026-09-28.