A Transformer-based conditional generative model augments skeleton action datasets by synthesizing high-fidelity sequences, improving recognition accuracy in few-shot and full-data regimes on HumanAct12 and NTU-VIBE.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2representative citing papers
RK4 at 80 function evaluations matches Euler at 200 in sliced Wasserstein quality for flow matching sampling, with the adaptive solver concentrating steps near t=1 due to stiffening velocity fields.
citing papers explorer
-
Generative Data Augmentation for Skeleton Action Recognition
A Transformer-based conditional generative model augments skeleton action datasets by synthesizing high-fidelity sequences, improving recognition accuracy in few-shot and full-data regimes on HumanAct12 and NTU-VIBE.
-
From Euler to Dormand-Prince: ODE Solvers for Flow Matching Generative Models
RK4 at 80 function evaluations matches Euler at 200 in sliced Wasserstein quality for flow matching sampling, with the adaptive solver concentrating steps near t=1 due to stiffening velocity fields.