본문 바로가기

인공지능 공부

 (45)
BOA (Bilevel Online Adaptation) for Out-of-Domain Human Mesh Reconstruction 읽기 https://sites.google.com/view/dynaboa/boa-cvpr-2021 Out-of-domain Human Mesh Reconstruction - BOA (CVPR 2021) Bilevel Online Adaptation for Out-of-Domain Human Mesh Reconstruction Shanyan Guan*, Jingwei Xu*, Yunbo Wang†, Bingbing Ni†, Xiaokang Yang MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University, China [Code] [Paper] [Supp. sites.google.com CVPR 2021에 발표된 논문인 ..
PIFuHD: Multi-Level Pixel-Aligned Implicit Function for High-Resolution 3D Human Digitization CVPR 2020 (Oral Presentation) https://shunsukesaito.github.io/PIFuHD/ Abstract 최근 이미지 기반의 3D 휴먼 생성 기법이 딥 뉴럴네트워크의 발전에 힘입어 많은 기술적 발전을 이룩했다. 물론 기존의 방식들도 실 세계와 같은 상황에 적용 가능한 잠재성을 가지고 있긴 하지만, 대부분 입력 이미지가 표현하고있는 디테일을 놓치는 결과를 생성한다. 우리는 이 한계가 두가지 대조되는 두가지 요구조건에 의해 생성된다고 판단했다. 첫 번째 조건은 전반적 형상을 정확하게 예측해야 한다는 부분과 세부적인 형상을 모사해야 한다는 부분이다. 이 두가지 조건을 모두 충족하는 것은 기존의 하드웨어의 메모리의 한계로 인해 현재로서는 불가능하다. 그러다보니 기존의 기술들은..
VNect: Real-time 3D Human Pose Estimation with a Single RGBCamera TOG 2017 http://vcai.mpi-inf.mpg.de/projects/VNect/ VNect: Real-time 3D Human Pose Estimation with a Single RGB Camera, SIGGRAPH 2017 Update: The dataset is now available for download Update: Live demo of VNect at CVPR17! ( More details ) --> Update: For errata regarding performance on MPI-INF-3DHP test set, please refer here. Update: VNect Demo Model and Code Available For Download as a vcai.mp..
Mesh Graphormer https://paperswithcode.com/paper/mesh-graphormer#code Papers with Code - Mesh Graphormer 🏆 SOTA for 3D Hand Pose Estimation on FreiHAND (PA-MPJPE metric) paperswithcode.com Abstract 우리는 Mesh Graphormer 라고 명명한 graph-convolution-reinforced transformer 를 제안한다. Mesh Graphormer는 하나의 이미지로부터 3D human pose and mesh reconstruction을 수행한다. 최근 transformer 방식과 graph convolutional neural networks(GCNMNs) 들은 휴..
VIBE: Video Inference for Human Body Pose and Shape Estimation https://github.com/mkocabas/VIBE GitHub - mkocabas/VIBE: Official implementation of CVPR2020 paper "VIBE: Video Inference for Human Body Pose and Shape Estimatio Official implementation of CVPR2020 paper "VIBE: Video Inference for Human Body Pose and Shape Estimation" - GitHub - mkocabas/VIBE: Official implementation of CVPR2020 paper "VIBE: ... github.com CVPR 2020 논문 •Abstract •사람의 모션은 행동을 이해하..
PyTorch Image Models(timm) 퀵 가이드 https://towardsdatascience.com/getting-started-with-pytorch-image-models-timm-a-practitioners-guide-4e77b4bf9055 Getting Started with PyTorch Image Models (timm): a practitioner’s guide The purpose of this guide is to explore PyTorch Image Models (timm) from a practitioner’s point of view, for use in custom scripts. towardsdatascience.com 원문을 공부할겸 해석해본다. PyTorch Image Models(timm)은 이미지 모델, 옵티마이져..
Motion Retargetting based on Dilated Convolutions andSkeleton-specific Loss Functions https://github.com/medialab-ku/retargetting-tdcn GitHub - medialab-ku/retargetting-tdcn: PyTorch implementation of the EG 2020 paper: Motion Retargetting based on Dilated Convol PyTorch implementation of the EG 2020 paper: Motion Retargetting based on Dilated Convolutions and Skeleton-specific Loss Functions - GitHub - medialab-ku/retargetting-tdcn: PyTorch implementation ... github.com @article..
TransPose: Real-time 3D Human Translation and Pose Estimation with Six Inertial Sensors https://xinyu-yi.github.io/TransPose/ TransPose Xinyu Yi1, Yuxiao Zhou1, Feng Xu1 1Tsinghua University Accepted by SIGGRAPH 2021 xinyu-yi.github.io @article{TransPoseSIGGRAPH2021, author = {Yi, Xinyu and Zhou, Yuxiao and Xu, Feng}, title = {TransPose: Real-time 3D Human Translation and Pose Estimation with Six Inertial Sensors}, journal = {ACM Transactions on Graphics}, year = {2021}, month = {0..

티스토리툴바