Image Captioning: Transforming Objects into Words

TL;DR

This paper introduces the Object Relation Transformer, which enhances image captioning by explicitly modeling spatial relationships between objects using geometric attention, resulting in improved performance on standard benchmarks.

Contribution

The paper presents a novel geometric attention mechanism within a transformer architecture for image captioning, explicitly modeling object spatial relations.

Findings

Improved captioning metrics on MS-COCO dataset

Geometric attention enhances understanding of object relationships

Quantitative and qualitative analysis confirms effectiveness

Abstract

Image captioning models typically follow an encoder-decoder architecture which uses abstract image feature vectors as input to the encoder. One of the most successful algorithms uses feature vectors extracted from the region proposals obtained from an object detector. In this work we introduce the Object Relation Transformer, that builds upon this approach by explicitly incorporating information about the spatial relationship between input detected objects through geometric attention. Quantitative and qualitative results demonstrate the importance of such geometric attention for image captioning, leading to improvements on all common captioning metrics on the MS-COCO dataset.