ICC: Quantifying Image Caption Concreteness for Multimodal Dataset Curation.
Moran YanukaMorris AlperHadar Averbuch-ElorRaja GiryesPublished in: CoRR (2024)
Keyphrases
- image data
- single image
- image content
- multiscale
- template matching
- image features
- image analysis
- input image
- edge detection
- region of interest
- image representation
- image dataset
- image segmentation
- image classification
- image set
- bounding box
- test images
- image retrieval
- image pixels
- pixel values
- edge map
- million images
- image regions
- segmentation method
- caption text
- complex background
- spatial information
- hough transform
- visual features