Connecting modalities: Semi-supervised segmentation and annotation of images using unaligned text corpora.

Richard Socher, Li Fei-Fei

Browse the full CVPR paper archive.