Google llc (20240257550). READING ORDER WITH POINTER TRANSFORMER NETWORKS simplified abstract
Contents
READING ORDER WITH POINTER TRANSFORMER NETWORKS
Organization Name
Inventor(s)
Diego Martin Arroyo of Zurich (CH)
READING ORDER WITH POINTER TRANSFORMER NETWORKS - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240257550 titled 'READING ORDER WITH POINTER TRANSFORMER NETWORKS
The method described in the abstract involves processing an image of a document with multiple layout components to predict the reading order based on textual and visual information using a self-attention encoder/decoder.
- Receives an image representing a document with various layout components
- Identifies textual information associated with the layout components
- Identifies visual information associated with the layout components
- Combines textual and visual information
- Predicts the reading order using a self-attention encoder/decoder
Potential Applications: - Document processing and organization - Automated reading order prediction for improved document understanding
Problems Solved: - Streamlining document analysis and comprehension - Enhancing efficiency in processing textual and visual information in documents
Benefits: - Improved document processing accuracy - Enhanced reading order prediction for better document understanding
Commercial Applications: Title: Automated Document Analysis and Reading Order Prediction Technology Description: This technology can be utilized in industries such as legal, academic, and administrative sectors for efficient document processing and organization, leading to improved productivity and accuracy.
Questions about the technology: 1. How does this technology improve document analysis processes? - This technology enhances document analysis by combining textual and visual information to predict the reading order, improving overall document understanding. 2. What are the potential applications of automated reading order prediction in various industries? - Automated reading order prediction can be beneficial in industries such as legal, academic, and administrative sectors for streamlining document processing and organization.
Original Abstract Submitted
a method including receiving an image representing a document including a plurality of layout components, identifying textual information associated with the plurality of layout components, identifying visual information associated with the plurality of layout components, combining the textual information with the visual information, and predicting a reading order of the plurality of layout components based on the combined textual information and visual information using a self-attention encoder/decoder.
