Democratizing Contrastive Language-Image Pre-training: A CLIP Benchmark of Data, Model, and Supervision.
Yufeng CuiLichen ZhaoFeng LiangYangguang LiJing ShaoPublished in: CoRR (2022)
Keyphrases
- data model
- image data
- single image
- multiscale
- spatial information
- input image
- image segmentation
- image representation
- image classification
- image features
- image analysis
- image content
- image retrieval
- image pixels
- low level
- template matching
- test images
- region of interest
- segmentation method
- edge detection
- high resolution
- segmentation algorithm
- feature points
- keypoints
- image matching
- active learning
- image collections
- pixel values
- image set
- berkeley segmentation dataset
- optical flow
- multiresolution
- similarity measure
- high level