哈尔滨商业大学学报（自然科学版）2025，Vol.41Issue(1)：18-24,7.
网格辅助的调色板NeRF外观编辑模型
Palette NeRF appearance editing model with mesh assistance
摘要
Abstract
The current palette-based Neural Radiance Field(NeRF)models enabled intuitive and fine-grained appearance editing of the 3D scenes.However,the fine-grained local appearance editing was limited to target-level objects,and could not render high-quality views in real time.To address these issues,a palette NeRF appearance editing model with mesh-assisted was proposed.The model achieved appearance editing by decomposing the colors of the scene into highlight colors and diffuse colors formed by alpha blending of palette colors and baking the colors onto the surface of the scene's mesh model to speed up rendering.During local editing,the mesh model of the scene was used to assist in the selection of editing targets.The experimental results showed that the model was able to edit the appearance of 3D regions of arbitrary size and shape and render high-quality editing results in real time under public datasets of different data types.The model used meshes to accelerate the rendering of the palette NeRF and enabled pixel-level local appearance editing.关键词
调色板/神经辐射场/外观编辑/实时渲染/局部编辑Key words
palette/neural radiation field/appearance editing/real-time rendering/local editing分类
信息技术与安全科学引用本文复制引用
毕成龙..网格辅助的调色板NeRF外观编辑模型[J].哈尔滨商业大学学报（自然科学版）,2025,41(1):18-24,7.基金项目
安徽省自然科学基金(2308085MF218) （2308085MF218）
安徽省高等学校科学研究项目(2022AH040113) （2022AH040113）
