ART3D: 3D Gaussian Splatting for Text-Guided Artistic Scenes Generation

TL;DR

ART3D introduces a novel framework combining diffusion models and 3D Gaussian splatting to generate high-quality, consistent 3D artistic scenes from images, effectively bridging artistic and realistic representations.

Contribution

The paper presents an innovative method that integrates diffusion models with 3D Gaussian splatting and semantic transfer for improved 3D artistic scene generation.

Findings

Demonstrates superior content and structural consistency over existing methods.

Effectively bridges the gap between artistic and realistic images.

Enhances 3D scene generation with depth and semantic transfer techniques.

Abstract

In this paper, we explore the existing challenges in 3D artistic scene generation by introducing ART3D, a novel framework that combines diffusion models and 3D Gaussian splatting techniques. Our method effectively bridges the gap between artistic and realistic images through an innovative image semantic transfer algorithm. By leveraging depth information and an initial artistic image, we generate a point cloud map, addressing domain differences. Additionally, we propose a depth consistency module to enhance 3D scene consistency. Finally, the 3D scene serves as initial points for optimizing Gaussian splats. Experimental results demonstrate ART3D's superior performance in both content and structural consistency metrics when compared to existing methods. ART3D significantly advances the field of AI in art creation by providing an innovative solution for generating high-quality 3D artistic scenes.