LAViTeR: Learning Aligned Visual and Textual Representations Assisted by Image and Caption Generation.
Mohammad Abuzar ShaikhZhanghexuan JiDana MoukheiberSargur N. SrihariMingchen GaoPublished in: CoRR (2021)
Keyphrases
- low level
- auto annotation
- image data
- single image
- learning algorithm
- image representation
- mid level
- image content
- image classification
- image analysis
- image retrieval
- reinforcement learning
- multiscale
- input image
- edge detection
- visual cues
- high resolution
- visually similar
- visual appearance
- image segmentation
- web images
- visual data
- similarity measure
- learning process
- spatial relations
- segmentation algorithm
- feature points
- image descriptors
- keywords
- image features
- perceptual information
- multimedia