Blind Bleed-Through Removal for Scanned Historical Document Image With Conditional Random Fields.
Bin SunShutao LiXiao-Ping ZhangJun SunPublished in: IEEE Trans. Image Process. (2016)
Keyphrases
- document images
- conditional random fields
- historical documents
- graphical models
- document image analysis
- random fields
- crf model
- probabilistic model
- information extraction
- hidden markov models
- higher order
- generative model
- pairwise
- sequence labeling
- named entity recognition
- document analysis
- scanned document images
- structured prediction
- markov random field
- scanned documents
- scanned images
- superpixels
- semi markov
- feature selection
- binary images
- text lines
- bayesian networks
- handwritten documents
- image processing
- line extraction
- printed text
- machine learning