Image Captioners Are Scalable Vision Learners Too.
Michael TschannenManoj KumarAndreas SteinerXiaohua ZhaiNeil HoulsbyLucas BeyerPublished in: NeurIPS (2023)
Keyphrases
- multiscale
- single image
- image features
- input image
- image classification
- image content
- feature points
- image data
- image retrieval
- image analysis
- image pixels
- high resolution
- low level image processing
- template matching
- visual perception
- low level
- image segmentation
- region of interest
- vector field
- segmentation method
- image matching
- light source
- spatial information
- image regions
- learning experience
- computer vision
- test images
- image representation
- vision system
- lighting conditions
- image set
- image structure
- learning process
- web images
- feature vectors
- image processing