# Gaussian Scattering

GS-VTON is a 3D virtual fitting technology that utilizes Gaussian scattering (3DGS) as a 3D representation, enabling knowledge transfer from 2D virtual fitting models to 3D space while enhancing cross-view consistency. This technology integrates personalized information into pre-trained 2D virtual fitting models through low-rank adaptation (LoRA). Additionally, it introduces a character-aware 3DGS editing framework to maintain consistent cross-view appearances and high-quality 3D geometric structures during the editing process. GS-VTON demonstrates its high fidelity and advanced editing capabilities in 3D virtual fittings through extensive experiments and comparative analyses with existing methods, proving its effectiveness in this domain.

EgoGaussian is an advanced 3D scene reconstruction and dynamic object tracking technology. It can reconstruct 3D scenes and track the movement of objects dynamically using only RGB first-person perspective input. This technology leverages the unique discrete properties of Gaussian scattering to segment dynamic interactions from the background. Through a piece-wise online learning process, it utilizes the dynamic characteristics of human activities to reconstruct the evolution of the scene in chronological order and track the movement of rigid objects. EgoGaussian outperforms previous NeRF and dynamic Gaussian methods in wild video challenges and delivers exceptional quality in reconstructed models.

Human101 is a framework for quickly reconstructing human figures from a single view. It can train a 3D Gaussian model within 100 seconds and render images at a resolution of 1024 and above 60FPS without pre-saving the Gaussian properties of each frame. The Human101 pipeline is as follows: first, extract 2D human pose from the single-view video. Then, use the pose to drive a 3D simulator to generate corresponding 3D skeletal animation. Finally, construct a time-varying 3D Gaussian model based on the animation for real-time rendering.