Efficient use of training data in the n-tuple recognition method

Tarling, Roland and Rohwer, Richard (1993). Efficient use of training data in the n-tuple recognition method. Electronics letters, 29 (24), pp. 2093-2094.

Abstract

A simple technique is presented for improving the robustness of the n-tuple recognition method against inauspicious choices of architectural parameters, guarding against the saturation problem, and improving the utilisation of small data sets. Experiments are reported which confirm that the method significantly improves performance and reduces saturation in character recognition problems.

Divisions: Aston University (General)
Uncontrolled Keywords: character recognition,pattern recognition training data,n-tuple recognition method,robustness,architectural parameters,saturation problem
Published Date: 1993-11-25
Authors: Tarling, Roland
Rohwer, Richard

Download

Item under embargo

Export / Share Citation


Statistics

Additional statistics for this record