Splats in Splats: Robust and Effective 3D Steganography towards Gaussian Splatting

The recent introduction of UniLight proposes a unified representation for lighting, addressing the complexities of lighting in images. This innovative approach integrates various modalities, including text, images, and environment maps, into a shared latent space, enhancing the understanding and representation of lighting effects in visual content.

ReCamDriving has been introduced as a novel framework for generating camera-controlled video trajectories without the use of LiDAR, relying instead on dense 3D Gaussian Splatting (3DGS) renderings for enhanced geometric guidance. This approach aims to overcome limitations of existing methods that struggle with complex artifacts or sparse data. The framework employs a two-stage training process to refine camera control and improve video generation accuracy.

The introduction of FantasyStyle marks a significant advancement in 3D Gaussian Splatting (3DGS) by addressing challenges in style transfer, particularly multi-view inconsistency and content leakage. This framework utilizes diffusion model distillation to enhance cross-view consistency and control stylization, aiming to improve the quality of generated 3D content.