Inserting Faces inside Captions: Image Captioning with Attention Guided Merging

TL;DR

This paper introduces AstroCaptions, a new dataset for image captioning focused on recognizing public figures, and proposes a novel post-processing method that improves caption quality by inserting identified people's names using explainable AI and vision-language grounding.

Contribution

The work presents a new dataset and a novel post-processing approach that enhances image captioning by accurately inserting people's names, reducing biases and hallucinations.

Findings

Up to 93.2% of detected persons are correctly inserted into captions.

Significant improvements in BLEU, ROUGE, CIDEr, and METEOR scores.

Reduction in caption hallucinations.

Abstract

Image captioning models are widely used to describe recent and archived pictures with the objective of improving their accessibility and retrieval. Yet, these approaches tend to be inefficient and biased at retrieving people's names. In this work we introduce AstroCaptions, a dataset for the image captioning task. This dataset specifically contains thousands of public fig-ures that are complex to identify for a traditional model. We also propose a novel post-processing method to insert identified people's names inside the caption using explainable AI tools and the grounding capabilities of vi-sion-language models. The results obtained with this method show signifi-cant improvements of captions quality and a potential of reducing halluci-nations. Up to 93.2% of the persons detected can be inserted in the image captions leading to improvements in the BLEU, ROUGE, CIDEr and METEOR scores of each captioning model.