Neural Articulated Radiance Field

TL;DR

Neural Articulated Radiance Field (NARF) introduces a deformable 3D representation for articulated objects that learns from images, capturing pose-dependent appearance changes efficiently without requiring 3D shape supervision.

Contribution

NARF is a novel implicit representation that models pose-dependent appearance of articulated objects using only rigid transformations, enabling training from images with pose annotations.

Findings

Efficiently models pose-dependent appearance changes.

Generalizes well to novel object poses.

Does not require 3D shape supervision.

Abstract

We present Neural Articulated Radiance Field (NARF), a novel deformable 3D representation for articulated objects learned from images. While recent advances in 3D implicit representation have made it possible to learn models of complex objects, learning pose-controllable representations of articulated objects remains a challenge, as current methods require 3D shape supervision and are unable to render appearance. In formulating an implicit representation of 3D articulated objects, our method considers only the rigid transformation of the most relevant object part in solving for the radiance field at each 3D location. In this way, the proposed method represents pose-dependent changes without significantly increasing the computational complexity. NARF is fully differentiable and can be trained from images with pose annotations. Moreover, through the use of an autoencoder, it can learn appearance variations over multiple instances of an object class. Experiments show that the proposed method is efficient and can generalize well to novel poses. The code is available for research purposes at https://github.com/nogu-atsu/NARF