Abstract
Existing image captioning models do not generalize well to out-of-domain images containing novel scenes or objects. This limitation severely hinders the use of these models in real world applications dealing with images in the wild. We address this problem using a flexible approach that enables existing deep captioning architectures to take advantage of image taggers at test time, without re-training. Our method uses constrained beam search to force the inclusion of selected tag words in the output, and fixed, pretrained word embeddings to facilitate vocabulary expansion to previously unseen tag words. Using this approach we achieve state of the art results for out-of-domain captioning on MSCOCO (and improved results for in-domain captioning). Perhaps surprisingly, our results significantly outperform approaches that incorporate the same tag predictions into the learning algorithm. We also show that we can significantly improve the quality of generated ImageNet captions by leveraging ground-truth labels.
Original language | English |
---|---|
Title of host publication | Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing |
Subtitle of host publication | EMNLP 2017 |
Editors | Martha Palmer, Rebecca Hwa, Sebastian Riedel |
Place of Publication | Stroudsburg, PA |
Publisher | Association for Computational Linguistics |
Pages | 936-945 |
Number of pages | 10 |
ISBN (Electronic) | 9781945626838 |
DOIs | |
Publication status | Published - 2017 |
Event | Conference on Empirical Methods in Natural Language Processing (2017) - Copenhagen, Denmark Duration: 9 Sept 2017 → 11 Sept 2017 |
Conference
Conference | Conference on Empirical Methods in Natural Language Processing (2017) |
---|---|
Abbreviated title | EMNLP 2017 |
Country/Territory | Denmark |
City | Copenhagen |
Period | 9/09/17 → 11/09/17 |