Spatial Cross-Attention for Transformer-Based Image Captioning.
Khoa Anh NgoKyuhong ShimByonghyo ShimPublished in: ICASSP (2023)
Keyphrases
- spatial information
- input image
- image data
- image classification
- single image
- image features
- spatial relationships
- spatial distribution
- image content
- image representation
- multiscale
- spatial correlation
- template matching
- spatial frequency
- image pixels
- image analysis
- image retrieval
- spatial arrangement
- spatial relations
- similarity measure
- region of interest
- temporal continuity
- spatial and temporal
- image matching
- image regions
- segmentation method
- feature points
- edge detection
- spatio temporal
- image collections
- super resolution
- high resolution
- spatial location
- spatial layout
- video sequences
- image segmentation