Relation constraint self-attention for image captioning.
Junzhong JiMingzhan WangXiaodan ZhangMinglong LeiLiangqiong QuPublished in: Neurocomputing (2022)
Keyphrases
- image data
- single image
- image content
- multiscale
- image retrieval
- edge detection
- input image
- image representation
- image features
- image classification
- template matching
- image analysis
- image pixels
- high resolution
- test images
- similarity measure
- image noise
- keypoints
- segmentation method
- post processing
- segmentation algorithm
- image matching
- vector field
- lighting conditions
- image collections
- aerial images
- image segmentation
- image regions
- image reconstruction
- natural images
- multi view
- color images