Pixel-Level Segmentation of Handwritten and Printed Texts in Document Images with Deep Learning.
Danjing LiuSun ZhouPublished in: CISP-BMEI (2023)
Keyphrases
- pixel level
- document images
- deep learning
- printed text
- optical character recognition
- text lines
- scanned documents
- handwritten documents
- unsupervised learning
- image fusion
- character recognition
- superpixels
- foreground objects
- machine learning
- multiscale
- segmentation algorithm
- natural images
- object recognition
- image segmentation
- feature selection