A Hierarchical Approach for Generating Descriptive Image Paragraphs

Abstract

Recent progress on image captioning has made it possible to generate novel sentences describing images in natural language, but compressing an image into a single sentence can describe visual content in only coarse detail. While one new captioning approach, dense captioning, can potentially describe images in finer levels of detail by captioning many regions within an image, it in turn is unable to produce a coherent story for an image. In this paper we overcome these limitations by generating entire paragraphs for describing images, which can tell detailed, unified stories. We develop a model that decomposes both images and paragraphs into their constituent parts, detecting semantic regions in images and using a hierarchical recurrent neural network to reason about language. Linguistic analysis confirms the complexity of the paragraph generation task, and thorough experiments on a new dataset of image and paragraph pairs demonstrate the effectiveness of our approach.

CVPR 2017 spotlight

References in corpus (4)

• PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
• Deep Fragment Embeddings for Bidirectional Image Sentence Mapping
• Image Captioning with Semantic Attention
• A Hierarchical Neural Autoencoder for Paragraphs and Documents

Cited by in corpus (8)

• Less Is More: Picking Informative Frames for Video Captioning
• Learning to Disambiguate by Asking Discriminative Questions
• Learning to discover and localize visual objects with open vocabulary
• A Model for Auto-Programming for General Purposes
• On Architectures for Including Visual Information in Neural Language Models for Image Description
• Towards Diverse Paragraph Captioning for Untrimmed Videos
• Textual Description for Mathematical Equations
• Panoptic Narrative Grounding