PCSR refines noisy correspondences in cross-modal retrieval by dividing samples via pseudo-label consistency and applying adaptive optimization with robust losses and text replacement.
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
UNVERDICTED 1representative citing papers
citing papers explorer
-
PCSR: Pseudo-label Consistency-Guided Sample Refinement for Noisy Correspondence Learning
PCSR refines noisy correspondences in cross-modal retrieval by dividing samples via pseudo-label consistency and applying adaptive optimization with robust losses and text replacement.