Augmenters at SemEval-2023 Task 1: Enhancing CLIP in Handling Compositionality and Ambiguity for Zero-Shot Visual WSD through Prompt Augmentation and Text-To-Image Diffusion.
Jie LiYow-Ting ShiueYong-Siang ShihJonas GeipingPublished in: SemEval@ACL (2023)
Keyphrases
- web images
- low level
- word sense disambiguation
- multiscale
- image data
- image features
- single image
- input image
- image representation
- visual cues
- visual perception
- visual information
- image regions
- image classification
- image retrieval
- image segmentation
- visual appearance
- high level
- high resolution
- auto annotation
- text retrieval
- web image search
- image collections
- image content
- image search
- test images
- visually similar
- similarity measure
- visual data
- visual attributes
- natural language processing
- wordnet
- image annotation
- low level features
- key frames