computer-vision-2

3D Human Reconstruction

3260 papers • 126 benchmarks • 313 datasets

This task has no description! Would you like to contribute one?

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in 3d-human-reconstruction-2

Trend

Dataset

Best Model

Actions

AGORA

Expressive hands and faces dataset (EHF)

CAPE

Libraries

i

Use these libraries to find 3d-human-reconstruction-2 models and implementations

Datasets

Subtasks

No subtasks available.

Most implemented papers

PIFuHD: Multi-Level Pixel-Aligned Implicit Function for High-Resolution 3D Human Digitization

T. Simon, Jason M. Saragih, H. Joo, Shunsuke Saito•Tue Mar 31 2020

This work formulates a multi-level architecture that is end-to-end trainable and significantly outperforms existing state-of-the-art techniques on single image human shape reconstruction by fully leveraging 1k-resolution input images.

829

Content

CAPE

EHF

BEHAVE

Surreal

Expressive hands and faces dataset (EHF).

Dynamic FAUST

0

3D Human Mesh Regression With Dense Correspondence

Wanli Ouyang, P. Luo, Xiaogang Wang, Wang Zeng, Wentao Liu•Sun May 31 2020

A model-free 3D human mesh estimation framework, named DecoMR, is proposed, which explicitly establishes the dense correspondence between the mesh and the local image features in the UV space (i.e., IUV image), and a novel UV map that maintains most of the neighboring relations on the original mesh surface.

106 0

Paper Graph

Nerfies: Deformable Neural Radiance Fields

J. Barron, Ricardo Martin-Brualla, Dan B. Goldman, S. Seitz, Keunhong Park, Sofien Bouaziz, U. Sinha•Tue Nov 24 2020

This work presents the first method capable of photorealistically reconstructing deformable scenes using photos/videos captured casually from mobile phones and shows that it faithfully reconstructs non-rigidly deforming scenes and reproduces unseen views with high fidelity.

1355 0

Paper Graph

PyMAF: 3D Human Pose and Shape Regression with Pyramidal Mesh Alignment Feedback Loop

Wanli Ouyang, Limin Wang, Zhenan Sun, Yebin Liu, Hongwen Zhang, Yating Tian, Xinchi Zhou•Mon Mar 29 2021

A Pyramidal Mesh Alignment Feedback (PyMAF) loop to leverage a feature pyramid and rectify the predicted parameters explicitly based on the mesh-image alignment status in the authors' deep regressor is proposed.

357 0

Paper Graph

ICON: Implicit Clothed humans Obtained from Normals

Michael J. Black, Dimitrios Tzionas, Yuliang Xiu, Jinlong Yang•Wed Dec 15 2021

The proposed ICON (“Tmplicit Clothed humans Obtained from Normals”) enables avatar creation directly from video with personalized pose-dependent cloth deformation and outperforms the state of the art in reconstruction.

384 0

Paper Graph

SmoothNet: A Plug-and-Play Network for Refining Human Poses in Videos

Ailing Zeng, Xu Ju, Lei Yang, Jiefeng Li, Jianyi Wang, Qiang Xu•Sun Dec 26 2021

SmoothNet models the natural smoothness characteristics in body movements by learning the long-range temporal relations of every joint without considering the noisy correlations among joints, and improves the temporal smoothness of existing pose estimators significantly and enhances the estimation accuracy of those challenging frames as a side-effect.

98 0

Paper Graph

AnthroNet: Conditional Generation of Humans via Anthropometrics

Sujoy Ganguly, A. Crespi, Jonathan Hogins, Francesco Picetti, Shrinath Deshpande, Jonathan Leban, Soroosh Shahtalebi, Jay Patel, Peifeng Jing, Chunpu Wang, Charles Metze, Cameron Sun, Cera Laidlaw, J. Warren, Kathy Huynh, River Page, Salehe Erfanian Ebadi•Wed Sep 06 2023

This work presents a novel human body model formulated by an extensive set of anthropocentric measurements, which is capable of generating a wide range of human body shapes and poses and is the first of its kind to have been trained end-to-end using only synthetically generated data.

2 0

Paper Graph

Joint Reconstruction of 3D Human and Object via Contact-Based Refinement Transformer

Gyeongsik Moon, Hyeongjin Nam, Daniel Sungho Jung, Kyoung Mu Lee•Sat Apr 06 2024

This work presents a novel joint 3D human-object reconstruction method (CONTHO) that effectively exploits contact information between humans and objects and proposes a novel contact-based refinement Transformer that effectively aggregates human features and object features based on the estimated human-object contact.

20 0

Paper Graph

DeepHuman: 3D Human Reconstruction From a Single Image

Qionghai Dai, Zerong Zheng, Yebin Liu, Tao Yu, Yixuan Wei•Thu Mar 14 2019

DeepHuman, an image-guided volume-to-volume translation CNN for 3D human reconstruction from a single RGB image, leverages a dense semantic representation generated from SMPL model as an additional input to reduce the ambiguities associated with the reconstruction of invisible areas.

357 0

Paper Graph

Self-supervised Learning of Motion Capture

Ersin Yumer, H. Tung, Katerina Fragkiadaki, Hsiao-Wei Tung•Sun Dec 03 2017

This work proposes a learning based motion capture model that optimizes neural network weights that predict 3D shape and skeleton configurations given a monocular RGB video and shows that the proposed model improves with experience and converges to low-error solutions where previous optimization methods fail.

330 0

Paper Graph

Adding a benchmark result helps the community track progress.