DIFNet: Boosting Visual Information Flow for Image Captioning.
Mingrui WuXuying ZhangXiaoshuai SunYiyi ZhouChao ChenJiaxin GuXing SunRongrong JiPublished in: CVPR (2022)
Keyphrases
- information flow
- image data
- multiscale
- single image
- low level
- image features
- image retrieval
- flow analysis
- visual perception
- edge detection
- image classification
- image representation
- image content
- visual cues
- image collections
- visual appearance
- input image
- image segmentation
- visual effects
- image regions
- test images
- visual data
- social networks
- web images
- communication networks
- high resolution
- flow graph