[논문] Champ: Controllable and Consistent Human Image Animation with 3D Parametric Guidance
·
Video Generation
Champ: Controllable and Consistent Human Image Animation with 3D Parametric GuidanceIn this study, we introduce a methodology for human image animation by leveraging a 3D human parametric model within a latent diffusion framework to enhance shape alignment and motion guidance in curernt human generative techniques. The methodology utilizearxiv.org  Champ fudan-generative-vision.github.ioAbstract..
[논문] DreamWaltz: Make a Scene with Complex 3D Animatable Avatars
·
3D Human Reconstruction
DreamWaltz: Make a Scene with Complex 3D Animatable AvatarsWe present DreamWaltz, a novel framework for generating and animating complex 3D avatars given text guidance and parametric human body prior. While recent methods have shown encouraging results for text-to-3D generation of common objects, creating high-quaidea-research.github.io DreamWaltz: Make a Scene with Complex 3D Animatable Avatars..
[논문] POPDG: Popular 3D Dance Generation with PopDanceSet
·
Music-to-Dance
POPDG: Popular 3D Dance Generation with PopDanceSetGenerating dances that are both lifelike and well-aligned with music continues to be a challenging task in the cross-modal domain. This paper introduces PopDanceSet, the first dataset tailored to the preferences of young audiences, enabling the generationarxiv.org POPDG: Popular 3D Dance Generation with PopDanceSetZhenye Luo, Min Ren, Xuecai Hu,..
[논문] Controllable Group Choreography using Contrastive Diffusion
·
Music-to-Dance
Music-driven group choreography는 다양한 어려움이 있지만, 산업 분야에 응용할 분야가 상당히 많기 때문에 최근 들어 연구가 활발하게 진행되고 있는 분야라고 한다.하지만, 대부분의 최근 연구들은 high-fidelity long-term motion을 만들지 못하거나 controllable한 경험을 허용하지 않는다. 그래서 이 논문에서는 high-quality, customizable한 group dance generation을 consistency와 diversity를 효율적으로 제어함으로써 해결하고자 했다고 한다.구체적으로 diffusion-based 방식을 사용해서 input music에 coherent하면서 flexible한 dancer 수와 long-term group d..
[논문] EDGE: Editable Dance Generation from Music
·
Music-to-Dance
EDGE: Editable Dance Generation from MusicEDGE: Editable Dance Generation from Music, 2022.edge-dance.github.io EDGE: Editable Dance Generation From MusicDance is an important human art form, but creating new dances can be difficult and time-consuming. In this work, we introduce Editable Dance GEneration (EDGE), a state-of-the-art method for editable dance generation that is capable of creating ..
[논문] HumanSplat: Generalizable Single-Image Human 3DGS with Structure Priors
·
3D Human Reconstruction
HumanSplat: Generalizable Single-Image Human Gaussian Splatting with Structure PriorsDespite recent advancements in high-fidelity human reconstruction techniques, the requirements for densely captured images or time-consuming per-instance optimization significantly hinder their applications in broader scenarios. To tackle these issues, wearxiv.org HumanSplat: Generalizable Single-Image Human Gau..