Doc2Im: document to image conversion through self-attentive embedding.
Mithun Das GuptaPublished in: CoRR (2018)
Keyphrases
- eye tracking
- visual attention
- image features
- image data
- single image
- input image
- image content
- image regions
- image representation
- low level
- image retrieval
- image analysis
- image classification
- multiscale
- segmentation method
- edge detection
- feature points
- image pixels
- information retrieval
- vector space
- test images
- cover image
- document images
- image compression
- information retrieval systems
- high resolution
- keywords
- similarity measure
- image processing
- computer vision