# Computer Graphics

This method involves creating relightable radiance fields by leveraging priors extracted from 2D image diffusion models. The approach can convert multi-view data captured under single lighting conditions into a dataset with multi-illumination effects, represented through 3D Gaussian splats for relightable radiance fields. This method does not rely on precise geometric shapes and surface normals, making it well-suited for handling cluttered scenes with complex geometries and reflective BRDFs.

ZeST is an image material transfer technology co-developed by the research teams of the University of Oxford, Stability AI, and MIT CSAIL. It enables the transfer of materials from one image to another object without any prior training. ZeST supports single material transfer and can handle multi-material editing within a single image. Users can easily apply one material to multiple objects in an image. In addition, ZeST also supports fast image processing on devices, eliminating the dependence on cloud computing or server-side processing, significantly improving efficiency.

AnimatableDreamer is a framework for generating and reconstructing animatable non-rigid 3D models from single-eye videos. It is capable of creating non-rigid objects of different categories while adhering to the object movement extracted from the video. The key technology is the proposed canonical score distillation method, which simplifies the generation dimension from 4D to 3D, performs denoising across different frames in the video, and carries out the distillation process within a unique canonical space. This ensures consistent generation and realistic morphologies across different postures. With differentiable deformation, AnimatableDreamer elevates the 3D generator to 4D, providing a new perspective on the generation and reconstruction of non-rigid 3D models. Additionally, combining with the inductive knowledge of consistency diffusion models, canonical score distillation can regularize reconstruction from new perspectives, thereby enhancing the generative process in a closed loop. Extensive experiments demonstrate that this method can generate highly flexible 3D models guided by text from single-eye videos, while achieving superior reconstruction performance compared to typical non-rigid reconstruction methods.

Paint3D generates high-resolution, unlit, and diverse 2K UV texture maps for untextured 3D meshes, and performs conditional generation based on text or image inputs. It first generates a rough texture map by creating view-dependent images with a pre-trained 2D diffusion model that considers depth information and then performs multi-view texture fusion. Subsequently, it employs specialized UV completion and UVHD texture models to remove the lighting effects and fill in incomplete areas. Paint3D can produce semantic-consistent, unlit 2K UV textures, significantly enhancing the texture generation quality of untextured 3D objects.

HAAR is a text-input based generation model capable of generating realistic 3D hairstyles. It takes text prompts as input and generates 3D hairstyle assets ready for various computer graphics animation applications. Unlike current AI-based generation models, HAAR utilizes 3D hair strands as the foundation representation. It automatically annotates the generated synthetic hairstyle model using a 2D visual question answering system. We propose a text-guided generation method that uses a conditional diffusion model to generate guided hair strands in the latent hairstyle UV space and reconstructs a dense hairstyle containing hundreds of thousands of hair strands using a latent upsampling process given a textual description. The generated hairstyles can be rendered using existing computer graphics techniques.