Automatic image cropping based on bottom-up saliency and top-down semantics.
Jiang LinIvan V. BajicPublished in: PACRIM (2017)
Keyphrases
- attention mechanism
- image content
- image data
- figure ground
- visual attention
- visual saliency
- image analysis
- input image
- image features
- image representation
- single image
- saliency map
- saliency detection
- image segmentation
- multiscale
- image retrieval
- salient regions
- pixel level
- salient region detection
- high resolution
- segmentation method
- edge detection
- saliency model
- figure ground segmentation
- information content
- bounding box
- bayesian framework
- region of interest
- keypoints
- salient points
- image watermarking
- image editing
- digital images
- geometric attacks