A new benchmark degrades NTU-120 skeleton data three ways, shows degradation type strongly affects accuracy, and finds LogSigRNN overtakes DeGCN at 3 FPS once missing frames are interpolated.
Action Recognition for Human–Robot Teaming: Exploring Mutual Performance Monitoring Possibilities,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SHARDeg: A Benchmark for Skeletal Human Action Recognition in Degraded Scenarios
A new benchmark degrades NTU-120 skeleton data three ways, shows degradation type strongly affects accuracy, and finds LogSigRNN overtakes DeGCN at 3 FPS once missing frames are interpolated.