A Fast Fully Octave Convolutional Neural Network for Document Image Segmentation.
Ricardo Batista das Neves JuniorLuiz Felipe VerçosaDavid MacêdoByron Leite Dantas BezerraCleber ZanchettinPublished in: CoRR (2020)
Keyphrases
- xml documents
- convolutional neural network
- image segmentation
- structured documents
- semantic information
- face detection
- graph cuts
- boundary detection
- document collections
- information retrieval
- document images
- document classification
- gray level
- image segmentation algorithm
- document clustering
- active contours
- information retrieval systems
- unsupervised image segmentation
- markov random field
- image processing
- level set method
- neural network
- multiscale
- region growing
- retrieval systems
- filter bank
- image analysis
- text documents
- relevant documents
- vector space model
- keywords
- frequency band
- contour detection
- web documents
- probabilistic relaxation
- method for image segmentation
- expectation maximization