Text-DIAE: A Self-Supervised Degradation Invariant Autoencoder for Text Recognition and Document Enhancement.
Mohamed Ali SouibguiSanket BiswasAndrés MaflaAli Furkan BitenAlicia FornésYousri KessentiniJosep LladósLluís GómezDimosthenis KaratzasPublished in: AAAI (2023)
Keyphrases
- text recognition
- text lines
- document images
- optical character recognition
- viterbi algorithm
- hidden markov models
- text documents
- text regions
- information retrieval
- document classification
- keywords
- web documents
- image processing
- document clustering
- information retrieval systems
- n gram
- retrieval systems
- semantic information
- connected components
- character recognition
- markov model
- dynamic programming