FooDI-ML: a large multi-language dataset of food, drinks and groceries images and descriptions.
David Amat OlóndrizPonç Palau PuigdevallAdrià Salvador PalauPublished in: CoRR (2021)
Keyphrases
- image database
- image retrieval
- ground truth
- image analysis
- input image
- image features
- three dimensional
- image classification
- image data
- object recognition
- natural language descriptions
- image dataset
- image set
- multiple images
- image matching
- image registration
- images with ground truth
- rigid body
- image annotation
- image collections
- similarity measure
- million images
- language learning
- scale space
- illumination conditions
- image understanding
- annotated images
- image regions
- segmentation algorithm