PAEE: Parameter-Efficient and Data-Effective Image Captioning Model with Knowledge Prompter and Cross-Modal Representation Aligner.
Yunji TianZhiming LiuQuan ZouGeng ChenPublished in: APWeb-WAIM (2) (2023)
Keyphrases
- image data
- input data
- semantic space
- multiscale
- background knowledge
- data processing
- visual data
- image representation
- data points
- similarity measure
- cross modal
- multi modal
- data sources
- image retrieval
- image features
- low level
- feature vectors
- spatial data
- image content
- image collections
- data analysis
- high level
- perceptual information