Towards Unique and Informative Captioning of Images

Abstract

Despite considerable progress, state of the art image captioning models produce generic captions, leaving out important image details. Furthermore, these systems may even misrepresent the image in order to produce a simpler caption consisting of common concepts. In this paper, we first analyze both modern captioning systems and evaluation metrics through empirical experiments to quantify these phenomena. We find that modern captioning systems return higher likelihoods for incorrect distractor sentences compared to ground truth captions, and that evaluation metrics like SPICE can be 'topped' using simple captioning systems relying on object detectors. Inspired by these observations, we design a new metric (SPICE-U) by introducing a notion of uniqueness over the concepts generated in a caption. We show that SPICE-U is better correlated with human judgements compared to SPICE, and effectively captures notions of diversity and descriptiveness. Finally, we also demonstrate a general technique to improve any existing captioning model -by using mutual information as a re-ranking objective during decoding. Empirically, this results in more unique and informative captions, and improves three different state-of-the-art models on SPICE-U as well as average score over existing metrics. 3

Is An Image Worth Five Sentences? A New Look into Semantics for Image-Text Matching

“…A direct outcome is that current evaluation solely considers those 5 sentences as relevant to a single image for the ITM task. However, it is a known fact that in MSCOCO or Flickr30k there are many sentences that can perfectly describe a non paired image [32,42,43]. In other words, there are sentences (images) that are relevant to images (sentences) even though they are not defined as such in the retrieval ground truth.…”