What value do explicit high level concepts have in vision to language problems?

Abstract

Much of the recent progress in Vision-to-Language (V2L) problems has been achieved through a combination of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). This approach does not explicitly represent high-level semantic concepts, but rather seeks to progress directly from image features to text. We propose here a method of incorporating high-level concepts into the very successful CNN-RNN approach, and show that it achieves a significant improvement on the state-of-the-art performance in both image captioning and visual question answering. We also show that the same mechanism can be used to introduce external semantic information and that doing so further improves performance. In doing so we provide an analysis of the value of high level semantic information in V2L problems.

Accepted to IEEE Conf. Computer Vision and Pattern Recognition 2016. Fixed title

References in corpus (18)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Sequence to Sequence Learning with Neural Networks
• Show, Attend and Tell: Neural Image Caption Generation with Visual Attention
• Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation
• Microsoft COCO Captions: Data Collection and Evaluation Server
• Unifying Visual-Semantic Embeddings with Multimodal Neural Language Models
• VQA: Visual Question Answering
• Deep Fragment Embeddings for Bidirectional Image Sentence Mapping
• Multiscale Combinatorial Grouping for Image Segmentation and Object Proposal Generation
• Are You Talking to a Machine? Dataset and Methods for Multilingual Image Question Answering
• Describing Videos by Exploiting Temporal Structure
• Ask Your Neurons: A Neural-based Approach to Answering Questions about Images
• Learning a Recurrent Visual Representation for Image Caption Generation
• Language Models for Image Captioning: The Quirks and What Works
• Aligning where to see and what to tell: image caption with region-based attention and scene factorization
• Guiding Long-Short Term Memory for Image Caption Generation
• Building a Large-scale Multimodal Knowledge Base System for Answering Visual Queries
• Hard to Cheat: A Turing Test based on Answering Questions about Images

Cited by in corpus (17)

• Explicit Knowledge-based Reasoning for Visual Question Answering
• Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering
• Knowing When to Look: Adaptive Attention via A Visual Sentinel for Image Captioning
• Ask Me Anything: Free-form Visual Question Answering Based on Knowledge from External Sources
• VQS: Linking Segmentations to Questions and Answers for Supervised Attention in VQA and Question-Focused Semantic Segmentation
• Semantic Compositional Networks for Visual Captioning
• Are You Talking to Me? Reasoned Visual Dialog Generation through Adversarial Learning
• Skeleton Key: Image Captioning by Skeleton-Attribute Decomposition
• Deep Compositional Captioning: Describing Novel Object Categories without Paired Training Data
• Tell-and-Answer: Towards Explainable Visual Question Answering using Attributes and Captions
• A Comprehensive Survey of Deep Learning for Image Captioning
• Disjoint Multi-task Learning between Heterogeneous Human-centric Tasks
• Answering Image Riddles using Vision and Reasoning through Probabilistic Soft Logic
• The VQA-Machine: Learning How to Use Existing Vision Algorithms to Answer New Questions
• Recurrent Models for Situation Recognition
• Topic-Guided Attention for Image Captioning
• Image2song: Song Retrieval via Bridging Image Content and Lyric Words