Spatial Relational Attention Using Fully Convolutional Networks for Image Caption Generation.
Teng JiangLiang GongYupu YangPublished in: Int. J. Comput. Intell. Appl. (2020)
Keyphrases
- spatial information
- spatial distribution
- image data
- image features
- single image
- input image
- multiscale
- template matching
- spatial frequency
- image classification
- image segmentation
- image content
- edge detection
- image analysis
- image representation
- spatial relationships
- image matching
- image retrieval
- segmentation method
- image collections
- spatio temporal
- spatial arrangement
- image pixels
- relative position
- network structure
- keypoints
- color distribution
- social networks
- spatial features
- relational graph
- caption text
- bounding box
- pixel values
- visual attention
- image regions
- visual features
- feature points
- image quality
- low level