作者: E.H. Ratzlaff
DOI: 10.1109/ICDAR.2003.1227737
关键词:
摘要: A framework of data organization methods and corresponding recognition results for UNIPEN databases is presented to enable the comparison from different isolated character recognizers. reproducible method splitting Train-R01/V07 into an array multi-writer omni-writer training testing pairs proposed. Recognition uncertainties are provided each pair, as well DevTest-R01/V02 subsets, using online scanning n-tuple recognizer. Several other published surveyed within this context. In sum, report provides reader multiple points reference useful comparing a number proposed that similarly allows private evaluation unpublished results.