Deep Fragment Embeddings for Bidirectional Image Sentence Mapping

TL;DR

This paper presents a novel bidirectional image-sentence retrieval model that embeds image objects and sentence relations into a shared space, improving retrieval accuracy and interpretability through fragment-level alignment.

Contribution

The model introduces a multi-modal embedding of image fragments and sentence fragments with an explicit alignment objective, enhancing retrieval performance and interpretability.

Findings

Significant improvement in image-sentence retrieval accuracy.

Explicit fragment alignment enhances interpretability.

Effective reasoning at both global and fragment levels.

Abstract

We introduce a model for bidirectional retrieval of images and sentences through a multi-modal embedding of visual and natural language data. Unlike previous models that directly map images or sentences into a common embedding space, our model works on a finer level and embeds fragments of images (objects) and fragments of sentences (typed dependency tree relations) into a common space. In addition to a ranking objective seen in previous work, this allows us to add a new fragment alignment objective that learns to directly associate these fragments across modalities. Extensive experimental evaluation shows that reasoning on both the global level of images and sentences and the finer level of their respective fragments significantly improves performance on image-sentence retrieval tasks. Additionally, our model provides interpretable predictions since the inferred inter-modal fragment alignment is explicit.