StyleNeRF: A Style-based 3D-Aware Generator for High-resolution Image Synthesis

TL;DR

StyleNeRF is a novel 3D-aware generative model that synthesizes high-resolution, multi-view consistent images with style and pose control, addressing previous limitations in detail, consistency, and controllability.

Contribution

It introduces a style-based generator integrated with NeRF, improving high-resolution image synthesis with 3D consistency, style control, and efficient training on unstructured 2D images.

Findings

Achieves high-resolution, multi-view consistent image synthesis

Enables style and pose manipulation with generalization to unseen views

Supports tasks like zoom, style mixing, and semantic editing

Abstract

We propose StyleNeRF, a 3D-aware generative model for photo-realistic high-resolution image synthesis with high multi-view consistency, which can be trained on unstructured 2D images. Existing approaches either cannot synthesize high-resolution images with fine details or yield noticeable 3D-inconsistent artifacts. In addition, many of them lack control over style attributes and explicit 3D camera poses. StyleNeRF integrates the neural radiance field (NeRF) into a style-based generator to tackle the aforementioned challenges, i.e., improving rendering efficiency and 3D consistency for high-resolution image generation. We perform volume rendering only to produce a low-resolution feature map and progressively apply upsampling in 2D to address the first issue. To mitigate the inconsistencies caused by 2D upsampling, we propose multiple designs, including a better upsampler and a new regularization loss. With these designs, StyleNeRF can synthesize high-resolution images at interactive rates while preserving 3D consistency at high quality. StyleNeRF also enables control of camera poses and different levels of styles, which can generalize to unseen views. It also supports challenging tasks, including zoom-in and-out, style mixing, inversion, and semantic editing.