Improving Handwritten OCR with Training Samples Generated by Glyph Conditional Denoising Diffusion Probabilistic Model.
Haisong DingBozhi LuanDongnan GuiKai ChenQiang HuoPublished in: CoRR (2023)
Keyphrases
- training samples
- denoising
- probabilistic model
- character recognition
- training set
- test sample
- nonlinear diffusion
- image denoising
- feature space
- learning algorithm
- character segmentation
- handwriting recognition
- optical character recognition
- bayesian networks
- hyperplane
- supervised learning
- training data
- decision boundary
- high dimensional
- number of training samples
- decision function
- face images
- document analysis
- noise removal
- sample set
- anisotropic diffusion
- printed documents
- machine vision
- database
- natural images
- base classifiers
- discriminative information
- image processing
- document images
- principal component analysis
- active learning
- image sequences
- computer vision
- machine learning