CVPR
2010
IEEE
15 years 2 days ago
2010
IEEE
We propose a semi-supervised model which segments and annotates images using very few labeled images and a large unaligned text corpus to relate image regions to text labels. Give...