Extending TrOCR for Text Localization-Free OCR of Full-Page Scanned Receipt Images.
Hongkuan ZhangEdward WhittakerIkuo KitagishiPublished in: CoRR (2022)
Keyphrases
- scanned images
- document images
- scanned documents
- printed documents
- text extraction
- image data
- image database
- three dimensional
- keywords
- line extraction
- optical character recognition
- image features
- page layout
- text regions
- input image
- ground truth
- text lines
- image registration
- image analysis
- object recognition
- document processing
- website
- scanned document images
- historical documents
- text recognition
- historical manuscripts
- image processing
- text detection
- text information
- image collections
- test images
- image classification
- document analysis
- localization method
- image annotation
- ocr systems
- video sequences
- image retrieval
- printed text
- edge detection
- lighting conditions
- digital camera
- textual information
- web images
- handwriting recognition