CoCa: Contrastive Captioners are Image-Text Foundation Models.
Jiahui YuZirui WangVijay VasudevanLegg YeungMojtaba SeyedhosseiniYonghui WuPublished in: Trans. Mach. Learn. Res. (2022)
Keyphrases
- image data
- image classification
- random fields
- input image
- multiscale
- image content
- probabilistic model
- image features
- image segmentation
- image representation
- image retrieval
- similarity measure
- bayesian framework
- segmentation method
- single image
- edge detection
- statistical model
- high resolution
- test images
- web images
- template matching
- parametric models
- pixel values
- image pixels
- feature points
- low level
- image matching
- image analysis
- text graphics
- text information
- textual and visual information
- image structure
- region of interest
- hough transform
- keypoints
- image quality