The authors introduce CSTBIR, a dataset of about 2 million sketch+text queries over 108K scene images, and STNET, a multimodal transformer that outperforms prior retrieval baselines on this new task.
K.; Sain, A.; Shah, P
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
-
Composite Sketch+Text Queries for Retrieving Objects with Elusive Names and Complex Interactions
The authors introduce CSTBIR, a dataset of about 2 million sketch+text queries over 108K scene images, and STNET, a multimodal transformer that outperforms prior retrieval baselines on this new task.