Title :
Transference of Learning Between Recognition Classes
Author_Institution :
Division of Computer Science, National Physical Laboratory, Teddington, Middlesex, England.
Abstract :
To demonstrate transference, experiments have been carried out in which the recognition of Highleyman´s handprinted numerals was a few percent more accurate when numerals and letters were used as a training set than when only numerals were used as a training set.
Keywords :
Character recognition; Distortion measurement; Learning systems; Machine learning; Pattern recognition; Prototypes; Rubber; Handprinted character recognition; learning machines; n-tuple methods; pattern recognition;
Journal_Title :
Computers, IEEE Transactions on
DOI :
10.1109/TC.1972.5008932
