Handwriting Character Recognision by using Fuzzy Logic

Authors : Enes Vardar, Kaplan Kaplan, H. Metin Ertunç

Pages : 71-77

View : 22 | Download : 10

Publication Date : 2017-10-01

Article Type : Research

Abstract :In Handwriting character recognision can be used to seek texts in big documents, take notes on tablet or decide whether or not internet user is a human or a computer in terms of Web security. In this study, a handwriting recognition system is studied by using fuzzy rules. The system includes 4 parts, namely image processing, feature extraction, fuzzification of the inputs, and defuzzification. In the first stage, image processing based on morphological operations are used to perform the handwriting recognisition under the same conditions. The feature extraction process is employed to find the total number of white pixels in each column. Then these pixel numbers are assigned to arrays. The next step is to find the local maximum and minimum values by considering this arrays as an increasing-decreasing mathematical function. Therefore, it is observed that the handwritten letters of these values are divided into various groups. In the next operation, fuzzy classification membership functions and rule tables of text groups are generated by using extracted feature data. For a better recognition perfromance, the letters group have to be known in order to use image fuzzy logic algorithm. Consequently, this group of letters was succesfully classified with fuzzy logic rules.
Keywords : Handwriting recognition, character recognisition, fuzzy logic approach, image processing algorithms