Turkish Times New Roman, Arial, and handwriting characters recognition by Neural Network Yapay Ṡinir Aǧlari i̇le Türkçe Times New Roman, Arial ve elyazisi Karakterleri̇ Tanima

Erdem O. A., Uzun E.

Journal of the Faculty of Engineering and Architecture of Gazi University, cilt.20, sa.1, ss.13-19, 2005 (SCI-Expanded) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 20 Sayı: 1
  • Basım Tarihi: 2005
  • Dergi Adı: Journal of the Faculty of Engineering and Architecture of Gazi University
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, TR DİZİN (ULAKBİM)
  • Sayfa Sayıları: ss.13-19
  • Anahtar Kelimeler: Artificial neural networks, Image processing, Pattern recognition, Recognation of handwriting
  • Gazi Üniversitesi Adresli: Evet

Özet

Artificial Neural Networks are input data systems which are developed for the cases in which are starting from the available data to disclose the unknown relationships and the algorithm or data processing system in which rules are not exactly known. In this study, the pattern recognition procedure which is an important application of neural network was considered. Times New Roman, Arial, and handwriting Turkish characters between A and Z with their lower cases were submitted to the networks as input vectors and the training procedures were realized according to these characters. Then, after applying character perceiving and recognition process to image which contains text written using these characters, image article was transformed to text. In this study a multilayer forward feeding neural network architecture and a back propagation training method were used. Also a character perceiving software was developed by using the C++ Builder software development environment.