Skip to main content

Towards Unique and Informative Captioning of Images

Author(s): Wang, Zeyu; Feng, Berthy; Narasimhan, Karthik; Russakovsky, Olga

To refer to this page use:
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 (Code is available at
Publication Date: 2020
Citation: Wang, Zeyu, Berthy Feng, Karthik Narasimhan, and Olga Russakovsky. "Towards Unique and Informative Captioning of Images." European Conference on Computer Vision (2020): pp. 629-644. doi:10.1007/978-3-030-58571-6_37
DOI: 10.1007/978-3-030-58571-6_37
ISSN: 0302-9743
EISSN: 1611-3349
Pages: 629 - 644
Type of Material: Conference Article
Journal/Proceeding Title: European Conference on Computer Vision
Version: Author's manuscript

Items in OAR@Princeton are protected by copyright, with all rights reserved, unless otherwise indicated.