Answers are grounded in this paper and cite their sources.
A CNN (cellular neural network) template for hole-filling is reported. There are many holes in Japanese characters and the CNN described can be used in addition to a connected-component-detector CNN for Japanese character recognition. It is expected that the hole-filler can probably also be used for other 2-D image recognition problems.>
No takes yet. Share an insight, caveat, or question.
Matsumoto et al. (1990) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: