GeoLayoutLM: Geometric Pre-training for Visual Information Extraction.
Chuwei LuoChangxu ChengQi ZhengCong YaoPublished in: CoRR (2023)
Keyphrases
- information extraction
- visual information
- free text
- conditional random fields
- visual features
- open domain
- visual perception
- named entity recognition
- machine learning
- natural language processing
- structured data
- low level
- text mining
- natural language text
- web mining
- precision and recall
- geometric structure
- question answering
- named entities
- training phase
- text summarization
- textual data
- training algorithm
- neural network
- visual data
- information retrieval
- feature selection
- word sense disambiguation
- support vector
- machine translation