Visual Spatial Description: Controlled Spatial-Oriented Image-to-Text Generation.
Yu ZhaoJianguo WeiZhichao LinYueheng SunMeishan ZhangMin ZhangPublished in: EMNLP (2022)
Keyphrases
- spatial information
- spatial relations
- spatial frequency
- spatial distribution
- spatial layout
- spatial relationships
- spatial arrangement
- image data
- image features
- multiscale
- spatial correlation
- image content
- single image
- visual information
- relative position
- template matching
- image description
- spatial data
- image retrieval
- text generation
- low level
- image classification
- keypoints
- visual perception
- input image
- web images
- visual data
- image collections
- image representation
- segmentation method
- visual effects
- visual attributes
- machine learning
- spatial configurations
- visual appearance
- color distribution
- visual features
- image analysis
- image segmentation