Pre-training image-language transformers for open-vocabulary tasks.
A. J. PiergiovanniWeicheng KuoAnelia AngelovaPublished in: CoRR (2022)
Keyphrases
- multiscale
- image data
- single image
- image content
- image analysis
- input image
- image collections
- image retrieval
- low level
- image segmentation
- feature points
- programming language
- image database
- image regions
- image matching
- template matching
- segmentation algorithm
- image representation
- image classification
- edge detection
- image features
- segmentation method
- training set
- region of interest
- similarity measure
- set of training images
- computer vision
- training examples
- high resolution
- natural language
- high level
- image structure
- bounding box
- image descriptors