The paper claims a first-place micro-gesture detection result from data augmentation and spatial-temporal attention, but its own table shows the winning F1 comes from the unmodified AdaTAD baseline, while the proposed method scores 33.44.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Online Micro-gesture Recognition Using Data Augmentation and Spatial-Temporal Attention
The paper claims a first-place micro-gesture detection result from data augmentation and spatial-temporal attention, but its own table shows the winning F1 comes from the unmodified AdaTAD baseline, while the proposed method scores 33.44.