A diagnostic study shows that two-stage HOI models fail differently across scene configurations like multi-person and rare interactions, revealing that aggregate benchmark accuracy does not imply robust visual reasoning.
Vi- sual compositional learning for human-object interaction de- tection
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A Study of Failure Modes in Two-Stage Human-Object Interaction Detection
A diagnostic study shows that two-stage HOI models fail differently across scene configurations like multi-person and rare interactions, revealing that aggregate benchmark accuracy does not imply robust visual reasoning.