SciOL and MuLMS-Img: Introducing A Large-Scale Multimodal Scientific Dataset and Models for Image-Text Tasks in the Scientific Domain.
Tim TarsiHeike AdelJan Hendrik MetzenDan ZhangMatteo FincoAnnemarie FriedrichPublished in: WACV (2024)
Keyphrases
- image data
- million images
- image dataset
- image features
- multiscale
- image retrieval
- image representation
- image content
- random fields
- single image
- image classification
- input image
- scientific data
- image segmentation
- web images
- bayesian framework
- image regions
- feature points
- segmentation method
- domain models
- image analysis
- low level
- scientific documents
- street view
- scientific literature
- probabilistic model
- region of interest
- data mining
- pixel values
- denoising
- photo collections
- object recognition
- image matching
- multiple modalities
- similarity measure
- transfer learning