PEMILIHAN CROSSOVER PADA ALGORITMA GENETIKA UNTUK PROGRAM APLIKASI PENGENALAN KARAKTER TULISAN TANGAN

Judah Suryaputra, Chairisni Lubis, Tri Sutrisno

Abstract


Handwriting recognition system using genetic algorithm is an Optical Character Recognition system which receives input in the form of handwritten image in scanned box and produces output in the form of characters from the handwriting. Writing can be uppercase, lowercase, or numbers. The designed system consists of five main processes: preprocessing the input image, vertical and horizontal line segmentation, line segmentation and character with automatic cropping, resizing template, and character recognition using genetic algorithm. Preprocessing of input image consists of grayscale process, and thresholding. Genetic algorithm is used to find characters from the character image obtained by comparing the image with the chromosome of the train data. To use the genetic algorithm method, given the process of resizing the template first so that the image size of the characters with the same template. This system has a success rate of character segmentation of 100% and success rate on character recognition with genetic algorithm of 89,027% with one point crossover, 90,43% with two point crossover, 90,72% with uniform crossover.

Full Text:

PDF

Refbacks

  • There are currently no refbacks.