S2TD: A Tree-Structured Decoder for Image Paragraph Captioning.
Yihui ShiYun LiuFangxiang FengRuifan LiZhanyu MaXiaojie WangPublished in: MMAsia (2021)
Keyphrases
- image representation
- image analysis
- image data
- image classification
- image retrieval
- image features
- input image
- multiscale
- image content
- image pixels
- region of interest
- segmentation method
- image segmentation
- template matching
- test images
- single image
- spatial information
- low level
- image collections
- learning algorithm
- hierarchical data structure
- structured data
- tree structure
- segmentation algorithm
- wavelet transform
- multiresolution
- pixel values
- aerial images
- successive approximation