Local-to-Global Semantic Supervised Learning for Image Captioning.
Juan WangYiping DuanXiaoming TaoJianhua LuPublished in: ICC (2020)
Keyphrases
- multiscale
- image data
- input image
- image content
- supervised learning
- template matching
- image classification
- single image
- image collections
- image retrieval
- image features
- image segmentation
- image analysis
- image pixels
- global features
- image representation
- hough transform
- global information
- semantically meaningful
- edge detection
- image matching
- reinforcement learning
- spatial information
- low level
- image regions
- keypoints
- segmentation method
- unsupervised learning
- high resolution
- pairwise
- training set
- video sequences
- keywords
- visual concepts
- semantic information
- machine learning