SIGIR
2006
ACM
15 years 9 months ago
2006
ACM
This paper employs ConceptNet, which covers a rich set of commonsense concepts, to retrieve images with text descriptions by focusing on spatial relationships. Evaluation on test ...