BlendGAN: Learning and Blending the Internal Distributions of Single Images by Spatial Image-Identity Conditioning.
Idan KligvasserTamar Rott ShahamNoa AlkobiTomer MichaeliPublished in: CoRR (2022)
Keyphrases
- input image
- image data
- image features
- image collections
- test images
- spatial information
- image retrieval
- image classification
- spatial frequency
- original images
- spatial relationships
- image analysis
- image regions
- high contrast
- image pixels
- image processing algorithms
- edge detection
- linear predictors
- segmentation method
- normalized correlation
- intensity values
- grey level
- image content
- spatial correlation
- spatial distribution
- imaging process
- pixel values
- image database
- segmentation algorithm
- human visual sensitivity
- image structure
- region of interest
- aerial images
- image quality
- image set
- contrast enhancement
- synthesized images
- image noise
- gray value
- color histogram
- unlabeled images
- feature points
- natural images
- web images
- false matches
- high resolution
- image matching
- bounding box
- quad trees
- labeled images
- digital imaging
- lighting conditions
- intensity distribution
- spatial location
- pixel intensities
- image representation
- sample images
- super resolution
- visual features
- single image
- vanishing points
- image segmentation
- image fusion
- object categories