3D Stylization

3D stylization aims to transfer artistic styles from images or text descriptions onto three-dimensional scenes, enabling the creation of visually appealing and diverse 3D content. Current research focuses on developing efficient and robust methods, often leveraging neural radiance fields (NeRFs) or Gaussian splatting representations, to achieve high-quality stylization while maintaining multi-view consistency and geometric accuracy. These advancements are driven by the need for improved tools in fields like game development, architectural visualization, and virtual reality, where the ability to quickly and easily stylize 3D models is highly valuable. The incorporation of text-based guidance further enhances the control and intuitiveness of the stylization process.