Action Recognition on NTU RGB+D

Metric: Accuracy (12 unseen classes) (higher is better)

LeaderboardDataset

Loading chart...

Results

Submit a result

Sort:

#	Model↕	Accuracy (12 unseen classes)▼	Extra Data	Paper	Date↕	Code
1	TDSM	56.03	No	TDSM: Triplet Diffusion for Skeleton-Text Matchi...	2024-11-16	Code
2	PGFA	55.99	No	Zero-shot Skeleton-based Action Recognition with...	2025-07-01	Code
3	STAR	45.1	No	Fine-Grained Side Information Guided Dual-Prompt...	2024-04-11	-
4	SA-DVAE	41.38	No	SA-DVAE: Improving Zero-Shot Skeleton-Based Acti...	2024-07-18	Code
5	PURLS	40.99	No	Part-aware Unified Representation of Language an...	2024-06-19	Code
6	SMIE	40.18	No	Zero-shot Skeleton-based Action Recognition via ...	2023-08-07	Code
7	SynSE	33.3	No	Syntactically Guided Generative Embeddings for Z...	2021-01-27	Code
8	CADA-VAE	28.96	No	Generalized Zero- and Few-Shot Learning via Alig...	2018-12-05	Code

#1TDSMSOTA
56.03
Accuracy (12 unseen classes)· 2024-11-16
TDSM: Triplet Diffusion for Skeleton-Text Matching in Zero-Shot Action Recognition Code
#2PGFA
55.99
Accuracy (12 unseen classes)· 2025-07-01
Zero-shot Skeleton-based Action Recognition with Prototype-guided Feature Alignment Code
#3STARSOTA
45.1
Accuracy (12 unseen classes)· 2024-04-11
Fine-Grained Side Information Guided Dual-Prompts for Zero-Shot Skeleton Action Recognition
#4SA-DVAE
41.38
Accuracy (12 unseen classes)· 2024-07-18
SA-DVAE: Improving Zero-Shot Skeleton-Based Action Recognition by Disentangled Variational Autoencoders Code
#5PURLS
40.99
Accuracy (12 unseen classes)· 2024-06-19
Part-aware Unified Representation of Language and Skeleton for Zero-shot Action Recognition Code
#6SMIESOTA
40.18
Accuracy (12 unseen classes)· 2023-08-07
Zero-shot Skeleton-based Action Recognition via Mutual Information Estimation and Maximization Code
#7SynSESOTA
33.3
Accuracy (12 unseen classes)· 2021-01-27
Syntactically Guided Generative Embeddings for Zero-Shot Skeleton Action Recognition Code
#8CADA-VAESOTA
28.96
Accuracy (12 unseen classes)· 2018-12-05
Generalized Zero- and Few-Shot Learning via Aligned Variational Autoencoders Code