English

Transform and Tell: Entity-Aware News Image Captioning

Computer Vision and Pattern Recognition 2020-06-16 v2 Computation and Language

Abstract

We propose an end-to-end model which generates captions for images embedded in news articles. News images present two key challenges: they rely on real-world knowledge, especially about named entities; and they typically have linguistically rich captions that include uncommon words. We address the first challenge by associating words in the caption with faces and objects in the image, via a multi-modal, multi-head attention mechanism. We tackle the second challenge with a state-of-the-art transformer language model that uses byte-pair-encoding to generate captions as a sequence of word parts. On the GoodNews dataset, our model outperforms the previous state of the art by a factor of four in CIDEr score (13 to 54). This performance gain comes from a unique combination of language models, word representation, image embeddings, face embeddings, object embeddings, and improvements in neural network design. We also introduce the NYTimes800k dataset which is 70% larger than GoodNews, has higher article quality, and includes the locations of images within articles as an additional contextual cue.

Keywords

Cite

@article{arxiv.2004.08070,
  title  = {Transform and Tell: Entity-Aware News Image Captioning},
  author = {Alasdair Tran and Alexander Mathews and Lexing Xie},
  journal= {arXiv preprint arXiv:2004.08070},
  year   = {2020}
}

Comments

Published in CVPR 2020. Code is available at https://github.com/alasdairtran/transform-and-tell and demo is available at https://transform-and-tell.ml

R2 v1 2026-06-23T14:54:51.034Z