Focus! Relevant and Sufficient Context Selection for News Image Captioning.
Mingyang ZhouGrace LuoAnna RohrbachZhou YuPublished in: EMNLP (Findings) (2022)
Keyphrases
- image data
- single image
- input image
- multiscale
- image retrieval
- image analysis
- image content
- image classification
- image segmentation
- spatial information
- test images
- image collections
- image features
- contextual information
- low level
- template matching
- region of interest
- grey level
- image representation
- edge detection
- hough transform
- computer vision
- feature points
- lighting conditions
- digital images
- image structure
- image pixels
- multiresolution
- image noise