Phrase Grounding by Soft-Label Chain Conditional Random Field

TL;DR

This paper introduces Soft-Label Chain CRFs for phrase grounding, modeling dependencies among image regions and handling multiple correct labels, achieving state-of-the-art results on Flickr30k Entities.

Contribution

It proposes a novel Soft-Label Chain CRF framework for phrase grounding, enabling end-to-end training and capturing entity dependencies effectively.

Findings

Achieved state-of-the-art performance on Flickr30k Entities.

Model benefits from entity dependency modeling and soft-label training.

Soft-Label Chain CRFs improve grounding accuracy.

Abstract

The phrase grounding task aims to ground each entity mention in a given caption of an image to a corresponding region in that image. Although there are clear dependencies between how different mentions of the same caption should be grounded, previous structured prediction methods that aim to capture such dependencies need to resort to approximate inference or non-differentiable losses. In this paper, we formulate phrase grounding as a sequence labeling task where we treat candidate regions as potential labels, and use neural chain Conditional Random Fields (CRFs) to model dependencies among regions for adjacent mentions. In contrast to standard sequence labeling tasks, the phrase grounding task is defined such that there may be multiple correct candidate regions. To address this multiplicity of gold labels, we define so-called Soft-Label Chain CRFs, and present an algorithm that enables convenient end-to-end training. Our method establishes a new state-of-the-art on phrase grounding on the Flickr30k Entities dataset. Analysis shows that our model benefits both from the entity dependencies captured by the CRF and from the soft-label training regime. Our code is available at \url{github.com/liujch1998/SoftLabelCCRF}