Semi-supervised cluster-and-label with feature based re-clustering to reduce noise in Thai document images.
N. PiroonsupSukree SinthupinyoPublished in: Knowl. Based Syst. (2015)
Keyphrases
- document images
- semi supervised
- constrained clustering
- clustering algorithm
- semi supervised clustering
- instance level constraints
- cluster membership
- arbitrary shape
- clustering approaches
- data clustering
- pair wise constraints
- hierarchical clustering
- k means
- clustering method
- cluster analysis
- cluster labels
- document image analysis
- data points
- unsupervised clustering
- document analysis
- semi supervised learning
- printed text
- inter cluster
- pairwise constraints
- optical character recognition
- scanned documents
- document image understanding
- spectral clustering
- noise level
- document image retrieval
- document processing
- median filter
- image binarization
- active learning
- cluster centers
- unlabeled data
- line extraction
- word spotting
- page layout
- mathematical formulas
- cluster ensemble
- noisy environments
- intra cluster
- page segmentation
- document clustering
- scanned document images
- labeled data
- noise reduction
- character recognition