sghong977 / Daily_AIML

Computer Vision, Deep Learning, 그외 MLOps 찍먹 등. 매일 새롭게 배운 것을 정리합니다.
0 stars 0 forks source link

[Survey List] NeuMan 후속 연구들 #31

Open sghong977 opened 1 year ago

sghong977 commented 1 year ago

노션이나 블로그나 컨플루언스 등등은 copy&paste할때 불편해서 여기다가 정리

generalization 관련

매번 사람 따로따로 학습하는거 덜어주는 논문. 이번에 이게 많이 보였다.

ActorsNeRF

YOTO

SHERF: Generalizable Human NeRF from a Single Image

Refining 3D Human Texture Estimation from a Single Image

속도 관련

GNeuVox

Real-time volumetric rendering of dynamic humans

Mesh Strikes Back: Fast and Efficient Human Reconstruction from RGB videos

누가 봐도 후속연구 & 비슷한 것

정말 정석적인 후속 연구 느낌이다 싶은 것들.

Detachable Novel Views Synthesis of Dynamic Scenes Using Distribution-Driven Neural Radiance Fields

HOSNeRF: object가 추가된 것

Vid2Avatar: 3D Avatar Reconstruction from Videos in the Wild via Self-supervised Scene Decomposition

sghong977 commented 1 year ago

인상적인 논문

"AvatarCraft: Transforming Text into Neural Human Avatars with Parameterized Shape and Pose Control"

NPC

"Total-Recon: Deformable Scene Reconstruction for Embodied View Synthesis"

sghong977 commented 1 year ago

Human에 포커스를 둔 것

특히 human pose에 포커스를 둔게 많은 듯

FlexNeRF: Photorealistic Free-viewpoint Rendering of Moving Humans from Sparse Views

AG3D

"PoseVocab: Learning Joint-structured Pose Embeddings for Human Avatar Modeling"

"MonoHuman: Animatable Human Neural Field from Monocular Video"

Structured 3D Features for Reconstructing Relightable and Animatable Avatars

InstantAvatar: Learning Avatars from Monocular Video in 60 Seconds

EMA: Efficient Meshy Neural Fields for Animatable Human Avatars

image

전신이 아닌 것

HandNeRF: Neural Radiance Fields for Animatable Interacting Hands

sghong977 commented 1 year ago

기타 논문들

"Hybrid Neural Rendering for Large-Scale Scenes with Motion Blur"

"RICO: Regularizing the Unobservable for Indoor Compositional Reconstruction"

Unsupervised Volumetric Animation

sghong977 commented 1 year ago

요약

해결하고 싶은 문제를 이미 누군가 비슷한 환경에서 해결했을 수 있어서, 현재 타겟으로 할 목표에 대해 아이디어 얻고자함.

핵심 내용 외 부가적으로 체크할 것