A MEASURE OF RECOGNITION DIFFICULTY FOR A CHARACTER IMAGE DATABASE

A character image database plays an important role not only in the dev elopment stage but also for the evaluation of a handwritten character recognition system. We propose a measure which tells the level of the recognition difficulty of a given character image database without the recognition process. We first define the entropy of each location of the image in the database. Then we obtain the measure by adding up the entropies of all locations in the image. We confirm that the proposed measure really reflects the relative recognition difficulties of data bases by using real databases. (C) 1998 Pattern Recognition Society. P ublished by Elsevier Science Ltd. All rights reserved.