This website requires JavaScript.

PaletteNeRF: Palette-based Color Editing for NeRFs

Qiling WuJianchao TanKun Xu
Dec 2022
摘要
Neural Radiance Field (NeRF) is a powerful tool to faithfully generate novelviews for scenes with only sparse captured images. Despite its strongcapability for representing 3D scenes and their appearance, its editing abilityis very limited. In this paper, we propose a simple but effective extension ofvanilla NeRF, named PaletteNeRF, to enable efficient color editing onNeRF-represented scenes. Motivated by recent palette-based image decompositionworks, we approximate each pixel color as a sum of palette colors modulated byadditive weights. Instead of predicting pixel colors as in vanilla NeRFs, ourmethod predicts additive weights. The underlying NeRF backbone could also bereplaced with more recent NeRF models such as KiloNeRF to achieve real-timeediting. Experimental results demonstrate that our method achieves efficient,view-consistent, and artifact-free color editing on a wide range ofNeRF-represented scenes.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答