Aligning Visual Regions and Textual Concepts for Semantic-Grounded Image Representations.
Fenglin LiuYuanxin LiuXuancheng RenXiaodong HeXu SunPublished in: NeurIPS (2019)
Keyphrases
- image representation
- visual concepts
- visual content
- visual and textual features
- image content
- visual features
- image classification
- low level features
- image features
- region segmentation
- multiscale
- semantic concepts
- bag of words
- visual representations
- visual information
- object recognition
- image search
- natural language
- image regions
- representation scheme
- low level
- scene categorization
- image retrieval
- textual descriptions
- feature representations
- visual words
- scene classification
- visual vocabulary
- web images
- computer vision
- cbir systems
- sparse coding
- semantic information
- text classification
- keywords
- high level
- image collections
- sparse representation
- spatial pyramid
- bag of features
- image annotation
- higher level
- scene recognition
- feature space
- spatial pyramid matching
- machine learning