A DNN-HMM-DNN Hybrid Model for Discovering Word-Like Units from Spoken Captions and Image Regions.

Liming Wang, Mark Hasegawa-Johnson

Browse the full Interspeech paper archive.