An autoencoder with joint-disentangled embeddings and vector-quantized temporal patches segments skeleton sequences into actions without labels, beating prior unsupervised methods on HuGaDB, LARa, and two of three BABEL subsets.
Emerg- ing properties in self-supervised vision transformers
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Skeleton Motion Words for Unsupervised Skeleton-Based Temporal Action Segmentation
An autoencoder with joint-disentangled embeddings and vector-quantized temporal patches segments skeleton sequences into actions without labels, beating prior unsupervised methods on HuGaDB, LARa, and two of three BABEL subsets.