The purposes of the current PE92 database project are twofold. One is to provide raw data to researchers so that they can concentrate their efforts primarily on the development of character recognition algorithms. The other is to provide a standard handwritten character data set to the perspective users as well as the developers so that they can evaluate and compare the performance of character recognition systems objectively. We collected 100 handwritten image sets of 2,350 Hanguel characters that correspond to the character set specified in Korean Standards KSC5601-1987 computer codes. We tried to collect as many writing styles as possible. The first 70 sets were generated by more than 500 different writers, and each of the remaining 30 sets was written by one person. Writers wrote down characters in the pre-specified boxes and the database was created by scanning the data sheets by an image scanner. The size of each image is 100100 pixels with 256 gray levels. Since each pixel needs one byte of memory, the size of the entire database PE92 turned out to be about 2.3 GB. Finally we obtained a raw data profile of PE92 by calculating various statistics of its image data.

This paper presents a system for recognition of on-line cursive Hangul (Korean characters) by means of DP matching of structural information. The penalty function has the following special features. In order to prevent short spurious strokes from causing large penalties, an input stroke is weighted by its length relative to other input strokes. In order to make use of pen-up and pen-down information, a penalty is incurred when 2 strokes of differing type (i.e. pen-up with pen-down) are matched. Finally, to reduce the chance of obtaining a suboptimal solution which can result from using the greedy algorithm in DP matching, we look-ahead an extra match. In a computer simulation we obtained a recognition rate of 92% for partially cursive characters and 89% for fully cursive characters. Furthermore, for both cases combined the correct character appears 98% of the time in the top 10 candidates. Thus we confirmed that the proposed algorithm is effective in recognizing cursive Hangul.