The field of generative modeling and animation is rapidly evolving, with a focus on creating more realistic and controllable digital avatars. Recent developments have led to significant improvements in photorealism, expression editing, and pose-dependent deformations. The use of diffusion models, Gaussian morphable models, and neural rendering has enabled the creation of high-fidelity digital avatars that can be used in various applications, including virtual try-on, animation, and video production. Noteworthy papers in this area include Face-MoGLE, which introduces a novel framework for controllable face generation, and Hyper Diffusion Avatars, which proposes a method for dynamic human avatar generation with high photorealism and realistic pose-dependent deformations. Additionally, papers such as Durian and Reconstruction and Reenactment Separated Method for Realistic Gaussian Head have made significant contributions to the field of portrait animation and 3D Gaussian head reconstruction.
Advances in Generative Modeling and Animation
Sources
Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation
Im2Haircut: Single-view Strand-based Hair Reconstruction for Human Avatars
GaussianGAN: Real-Time Photorealistic controllable Human Avatars
Towards High-Fidelity, Identity-Preserving Real-Time Makeup Transfer: Decoupling Style Generation
GRMM: Real-Time High-Fidelity Gaussian Morphable Head Model with Learned Residuals
Enhancing Zero-Shot Pedestrian Attribute Recognition with Synthetic Data Generation: A Comparative Study with Image-To-Image Diffusion Models
A Data-Centric Approach to Pedestrian Attribute Recognition: Synthetic Augmentation via Prompt-driven Diffusion Models
TeRA: Rethinking Text-driven Realistic 3D Avatar Generation
Hyper Diffusion Avatars: Dynamic Human Avatar Generation using Network Weight Space Diffusion
Durian: Dual Reference-guided Portrait Animation with Attribute Transfer
Reconstruction and Reenactment Separated Method for Realistic Gaussian Head
From Rigging to Waving: 3D-Guided Diffusion for Natural Animation of Hand-Drawn Characters