This paper discusses an application program design for character recognition with various font sizes to maximize the accuration. Common image processing usually uses grayscaling, thresholding, filtering, segmentation, and stretching. Those methods are purposed to renew images into simpler ones. Meanwhile, in this paper the character recognition uses fuzzy logic approach. By this approach, every single font on text can be recognized based on its own characteristic. The Output programs resulted in this study is a text that is ready to be processed further. Moreover, from the examination carried out, the one-by-one letter recognition system obtained accuration as much as 96.67% while the accuration obtained on the standard document recognition is only 96.13%
Manik N. I and Faisal. (2012). PERANCANGAN PROGRAM APLIKASI PENGENALAN TEKS MENGGUNAKAN FUZZY LOGIC. MatStat, 12 (1), 66-75.