XCiT: Cross-Covariance Image Transformers.
Alaaeldin AliHugo TouvronMathilde CaronPiotr BojanowskiMatthijs DouzeArmand JoulinIvan LaptevNatalia NeverovaGabriel SynnaeveJakob VerbeekHervé JégouPublished in: NeurIPS (2021)
Keyphrases
- image data
- input image
- multiscale
- single image
- image content
- template matching
- image features
- high resolution
- image analysis
- image classification
- image representation
- image retrieval
- segmentation method
- image segmentation
- grey level
- image pixels
- image collections
- region of interest
- test images
- aerial images
- image noise
- spatial information
- segmentation algorithm
- feature points
- edge detection
- image regions
- energy function
- lighting conditions
- post processing
- natural images
- low level