Editing Conditional Radiance Fields

TL;DR

This paper presents a novel method for user-guided editing of category-level neural radiance fields, enabling intuitive 3D modifications based on 2D scribbles, with applications demonstrated on shape datasets and real photographs.

Contribution

It introduces a modular conditional radiance field with a shape branch and a hybrid update strategy for efficient, accurate 3D editing based on coarse 2D user input.

Findings

Outperforms prior neural editing methods in accuracy and efficiency

Enables propagation of user edits to novel views and entire 3D regions

Successfully edits real photographs with consistent appearance and shape changes

Abstract

A neural radiance field (NeRF) is a scene model supporting high-quality view synthesis, optimized per scene. In this paper, we explore enabling user editing of a category-level NeRF - also known as a conditional radiance field - trained on a shape category. Specifically, we introduce a method for propagating coarse 2D user scribbles to the 3D space, to modify the color or shape of a local region. First, we propose a conditional radiance field that incorporates new modular network components, including a shape branch that is shared across object instances. Observing multiple instances of the same category, our model learns underlying part semantics without any supervision, thereby allowing the propagation of coarse 2D user scribbles to the entire 3D region (e.g., chair seat). Next, we propose a hybrid network update strategy that targets specific network components, which balances efficiency and accuracy. During user interaction, we formulate an optimization problem that both satisfies the user's constraints and preserves the original object structure. We demonstrate our approach on various editing tasks over three shape datasets and show that it outperforms prior neural editing approaches. Finally, we edit the appearance and shape of a real photograph and show that the edit propagates to extrapolated novel views.