A Fast Fully Octave Convolutional Neural Network for Document Image Segmentation.
Ricardo Batista das Neves JuniorLuiz Felipe VerçosaDavid MacedoByron Leite Dantas BezerraCleber ZanchettinPublished in: IJCNN (2020)
Keyphrases
- convolutional neural network
- image segmentation
- face detection
- document images
- information retrieval systems
- information retrieval
- document retrieval
- document collections
- multiscale
- web documents
- filter bank
- boundary detection
- document classification
- retrieval systems
- graph cuts
- markov random field
- unsupervised image segmentation
- image processing
- computer vision
- segmentation algorithm
- level set
- method for image segmentation
- text documents
- image analysis
- frequency band
- tf idf
- gray level
- radial basis
- document analysis
- keywords
- document representation
- segmentation method
- video sequences
- shape prior
- region growing
- k means
- neural network