Handwritten Digits Classification with Generalized Classifier Neural Network


ÖZYILDIRIM B. M., AVCI M.

Innovations in Intelligent Systems and Applications Conference (ASYU), Adana, Turkey, 4 - 06 October 2018, pp.187-189 identifier identifier

  • Publication Type: Conference Paper / Full Text
  • Volume:
  • Doi Number: 10.1109/asyu.2018.8553999
  • City: Adana
  • Country: Turkey
  • Page Numbers: pp.187-189
  • Çukurova University Affiliated: Yes

Abstract

Pattern recognition process is composed of three main steps: preprocessing, feature extraction and classification. Characteristic features and an efficient classifier are the key elements of successful recognition system. In this work, generalized classifier neural network; introduced as an efficient classifier; is used for classification of handwritten digits. Classification is implemented with MATLAB R2016a environment. Classification performances of multilayer perceptron, probabilistic neural network and generalized classifier neural network are compared and efficiency of generalized classifier neural network is explained. Generalized classifier neural network provides 99.9644% and 99.4086% classification performances on optical recognition handwritten data set and pen-based recognition handwritten data set, respectively