CG-NeRF: Conditional Generative Neural Radiance Fields

TL;DR

CG-NeRF introduces a conditional generative model that produces multi-view, diverse 3D-aware images based on user-provided inputs like images or text, improving fidelity and diversity over previous models.

Contribution

The paper presents a novel unified architecture for disentangling shape and appearance conditioned on various inputs, along with a pose-consistent diversity loss for multimodal output generation.

Findings

Maintains consistent image quality across different condition types

Achieves superior fidelity compared to existing models

Generates diverse multi-view images with detailed features

Abstract

While recent NeRF-based generative models achieve the generation of diverse 3D-aware images, these approaches have limitations when generating images that contain user-specified characteristics. In this paper, we propose a novel model, referred to as the conditional generative neural radiance fields (CG-NeRF), which can generate multi-view images reflecting extra input conditions such as images or texts. While preserving the common characteristics of a given input condition, the proposed model generates diverse images in fine detail. We propose: 1) a novel unified architecture which disentangles the shape and appearance from a condition given in various forms and 2) the pose-consistent diversity loss for generating multimodal outputs while maintaining consistency of the view. Experimental results show that the proposed method maintains consistent image quality on various condition types and achieves superior fidelity and diversity compared to existing NeRF-based generative models.