computer-vision-7

3D Human Dynamics

3260 papers • 126 benchmarks • 313 datasets

Image: Zhang et al

(Image credit: Papersgraph)

Benchmarks

These leaderboards are used to track progress in 3d-human-dynamics-7

Trend

Dataset

Best Model

Actions

No benchmarks available.

Libraries

i

Use these libraries to find 3d-human-dynamics-7 models and implementations

Datasets

InfiniteRep

Subtasks

No subtasks available.

Most implemented papers

Dynamical Variational Autoencoders: A Comprehensive Review

Xavier Alameda-Pineda, Xiaoyu Bie, Thomas Hueber, Laurent Girin, Simon Leglaive, Julien Diard•Thu Aug 27 2020

A general class of models called Dynamical Variational Autoencoders (DVAEs) are introduced that encompass a large subset of these temporal VAE extensions that not only model the latent space, but also model the temporal dependencies within a sequence of data vectors and/or corresponding latent vectors.

272

Content

0

Paper Graph

Learning 3D Human Dynamics From Video

Angjoo Kanazawa, Jitendra Malik, Jason Y. Zhang, Panna Felsen•Mon Dec 03 2018

The approach is designed so it can learn from videos with 2D pose annotations in a semi-supervised manner and obtain state-of-the-art performance on the 3D prediction task without any fine-tuning.

575 0

Paper Graph

Predicting 3D Human Dynamics From Video

Angjoo Kanazawa, Jitendra Malik, Jason Y. Zhang, Panna Felsen•Mon Aug 12 2019

This work presents perhaps the first approach for predicting a future 3D mesh model sequence of a person from past video input, and inspired by the success of autoregressive models in language modeling tasks, learns an intermediate latent space on which to predict the future.

115 0

Paper Graph

InterDiff: Generating 3D Human-Object Interactions with Physics-Informed Diffusion

Sirui Xu, Zhengyu Li, Yu-Xiong Wang, Liangyan Gui•Wed Aug 30 2023

This paper proposes InterDiff, a framework comprising two key steps: interaction diffusion, where a diffusion model is leverage to encode the distribution of future human-object interactions; and interaction correction, where a physics-informed predictor is introduced to correct denoised HOIs in a diffusion step.

179 0

Paper Graph

AutoAvatar: Autoregressive Neural Fields for Dynamic Avatar Modeling

J. Romero, M. Zollhöfer, Ping Tan, Shunsuke Saito, Ziqian Bai, Timur M. Bagautdinov•Thu Mar 24 2022

This work enables autoregressive modeling of implicit avatars and introduces the notion of articulated observer points, which relate implicit states to the explicit surface of a parametric human body model, and demonstrates that encoding implicit surfaces as a set of height fields defined on articulated observer Points leads to significantly better generalization compared to a latent representation.

4 0

Paper Graph

Adding a benchmark result helps the community track progress.