Page segmentation of historical document images with convolutional autoencoders.
Kai ChenMathias SeuretMarcus LiwickiJean HennebertRolf IngoldPublished in: ICDAR (2015)
Keyphrases
- page segmentation
- document images
- historical documents
- restricted boltzmann machine
- deep belief networks
- denoising
- document image analysis
- optical character recognition
- deep learning
- document image understanding
- document analysis
- document layout
- scanned documents
- image segmentation
- comparative evaluation
- document image retrieval
- image processing
- evaluation methods
- image enhancement
- printed text
- multiscale
- scanned document images