Human action generation

Yan et al. (2019) CSGN:

"When the dancer is stepping, jumping and spinning on the stage, attentions of all audiences are attracted by the streamof the fluent and graceful movements. Building a model that is capable of dancing is as fascinating a task as appreciating the performance itself. In this paper, we aim to generate long-duration human actions represented as skeleton sequences, e.g. those that cover the entirety of a dance, with hundreds of moves and countless possible combinations."

( Image credit: Convolutional Sequence Generation for Skeleton-Based Action Synthesis )

Papers

Recently Added Most Hyped Most Active Needs Verification Most Verified

Showing 1–10 of 13 papers

Title	Date	Tasks	Status	Hype
FinePhys: Fine-grained Human Action Generation by Explicitly Incorporating Physical Laws for Effective Skeletal Guidance	May 19, 2025	Action GenerationHuman action generation	—Unverified	0
FLAG3D: A 3D Fitness Activity Dataset with Language Instruction	Dec 9, 2022	Action GenerationAction Recognition	CodeCode Available	0
Action-conditioned On-demand Motion Generation	Jul 17, 2022	Contrastive LearningDecoder	CodeCode Available	0
MUGL: Large Scale Multi Person Conditional Action Generation with Locomotion	Oct 21, 2021	Action GenerationDiversity	CodeCode Available	1
Generative Adversarial Graph Convolutional Networks for Human Action Synthesis	Oct 21, 2021	Action GenerationDisentanglement	CodeCode Available	1
Action-Conditioned 3D Human Motion Synthesis with Transformer VAE	Apr 12, 2021	Action RecognitionDenoising	CodeCode Available	1
Action2Motion: Conditioned Generation of 3D Human Motions	Jul 30, 2020	Action GenerationHuman action generation	CodeCode Available	1
Structure-Aware Human-Action Generation	Jul 4, 2020	Action Generationgraph construction	CodeCode Available	1
Learning Diverse Stochastic Human-Action Generators by Learning Smooth Latent Transitions	Dec 21, 2019	Action GenerationDecoder	CodeCode Available	0
Convolutional Sequence Generation for Skeleton-Based Action Synthesis	Oct 27, 2019	Human action generation	—Unverified	0

Show:10 25 50

← PrevPage 1 of 2Next →

All datasets Human3.6M NTU RGB+D 2D NTU RGB+D NTU RGB+D 120 CMU Mocap HumanAct12 UESTC RGB-D

Benchmark Results

#	Model	Metric	Claimed	Verified	Status
1	Deep Video Generation, Prediction and Completion of Human Action Sequences	MMDa	0.42	—	Unverified
2	Learning Diverse Stochastic Human-Action Generators by Learning Smooth Latent Transitions	MMDa	0.2	—	Unverified
3	c-GAN	MMDa	0.16	—	Unverified
4	SA-GCN	MMDa	0.15	—	Unverified
5	Kinetic-GAN	MMDa	0.07	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	SkeletonGAN	MMDa (CS)	0.7	—	Unverified
2	c-SkeletonGAN	MMDa (CS)	0.34	—	Unverified
3	c-GAN	MMDa (CS)	0.33	—	Unverified
4	SA-GCN	MMDa (CS)	0.29	—	Unverified
5	Kinetic-GAN	MMDa (CS)	0.26	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	c-GAN	FID (CS)	27.48	—	Unverified
2	CSGN	FID (CS)	6.03	—	Unverified
3	Kinetic-GAN	FID (CS)	3.62	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	c-GAN	FID (CS)	54.4	—	Unverified
2	Kinetic-GAN	FID (CS)	5.97	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	ODMO	Accuracy	93.51	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	ODMO	Accuracy	97.81	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	ODMO	Accuracy	93.67	—	Unverified