VICTR: Visual Information Captured Text Representation for Text-to-Vision Multimodal Tasks.
Soyeon Caren HanSiqu LongSiwen LuoKunze WangJosiah PoonPublished in: COLING (2020)
Keyphrases
- visual information
- text representation
- textual information
- text documents
- visual features
- concept learning
- keywords
- low level
- statistical natural language processing
- text retrieval
- text categorization
- information filtering
- text mining
- text classification
- bag of words
- document representation
- index terms
- semantic information
- document clustering
- vector space model
- eye movements
- computer vision
- information retrieval
- vision system
- image processing
- word sense disambiguation
- text clustering
- machine learning
- information extraction
- domain knowledge
- user profiles
- query expansion
- natural language processing
- data analysis
- high level
- databases