Could Giant Pretrained Image Models Extract Universal Representations?
Yutong LinZe LiuZheng ZhangHan HuNanning ZhengStephen LinYue CaoPublished in: CoRR (2022)
Keyphrases
- multiscale
- single image
- image data
- random fields
- image features
- input image
- image segmentation
- image analysis
- bayesian framework
- image representation
- object models
- image classification
- shape properties
- parametric models
- image pixels
- image retrieval
- statistical model
- image matching
- low level
- object recognition
- hough transform
- image content
- vector field
- template matching
- surface model
- graph cuts
- high resolution
- image set
- visual effects
- probability models
- visual patterns
- probabilistic model
- image collections
- image database
- region of interest
- test images
- energy function