Deep visual-semantic alignments for generating image descriptions

TLDR

The paper proposes a model that generates natural language descriptions for images and their regions. The approach learns cross‑modal alignments with a CNN–RNN architecture and a multimodal embedding objective, then uses a multimodal recurrent network to generate novel descriptions of image regions. The model achieves state‑of‑the‑art retrieval performance on Flickr8K, Flickr30K, and MSCOCO, and its generated descriptions outperform retrieval baselines on full images and a new region‑level dataset.

Abstract

We present a model that generates natural language descriptions of images and their regions. Our approach leverages datasets of images and their sentence descriptions to learn about the inter-modal correspondences between language and visual data. Our alignment model is based on a novel combination of Convolutional Neural Networks over image regions, bidirectional Recurrent Neural Networks over sentences, and a structured objective that aligns the two modalities through a multimodal embedding. We then describe a Multimodal Recurrent Neural Network architecture that uses the inferred alignments to learn to generate novel descriptions of image regions. We demonstrate that our alignment model produces state of the art results in retrieval experiments on Flickr8K, Flickr30K and MSCOCO datasets. We then show that the generated descriptions significantly outperform retrieval baselines on both full images and on a new dataset of region-level annotations.