FIRE-CIR improves composed image retrieval accuracy on Fashion IQ by using generated visual questions for explicit attribute reasoning and candidate re-ranking instead of pure embedding similarity.
Improving composed image retrieval via contrastive learning with scal- ing positives and negatives
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
-
FIRE-CIR: Fine-grained Reasoning for Composed Fashion Image Retrieval
FIRE-CIR improves composed image retrieval accuracy on Fashion IQ by using generated visual questions for explicit attribute reasoning and candidate re-ranking instead of pure embedding similarity.