LLAFN-Generator: Learnable linear-attention with fast-normalization for large-scale image captioning.
Xiaobao YangXi TianJunsheng WuXiaochun YangSugang MaXinman QiZhiqiang HouPublished in: Comput. Vis. Image Underst. (2024)
Keyphrases
- input image
- image data
- image content
- single image
- image retrieval
- image analysis
- image classification
- spatial filters
- image features
- multiscale
- segmentation method
- image representation
- image pixels
- edge detection
- image collections
- template matching
- learning algorithm
- million images
- normalization method
- region of interest
- spatial information
- image segmentation
- similarity measure
- vector field
- image processing
- keypoints
- image structure
- differential operators
- feature points
- low level