UrbanCLIP: Learning Text-enhanced Urban Region Profiling with Contrastive Language-Image Pretraining from the Web.
Yibo YanHaomin WenSiru ZhongWei ChenHaodong ChenQingsong WenRoger ZimmermannYuxuan LiangPublished in: WWW (2024)
Keyphrases
- image data
- region of interest
- input image
- web images
- image segmentation
- image pixels
- image classification
- multiscale
- information retrieval
- image features
- image regions
- web documents
- single image
- language learning
- programming language
- image retrieval
- learning algorithm
- segmentation algorithm
- computational linguistics
- region segmentation
- image content
- low level
- natural language
- image representation
- feature points
- aerial images
- edge map
- textual data
- color images
- grey level
- learning process
- language generation