作者: Masoud Arabfard , Meisam Askari , Milad Asadi , Hosein Ebrahimpour-Komleh
DOI: 10.1007/978-3-642-27337-7_28
关键词:
摘要: In this paper we propose a system for recognition of isolated handwritten Persian characters. A novel method that uses derivation has been used feature extraction. Hamming network classification in system. is neural fully connected from input layer to all neuron output which calculate amount resemblance between patterns than training patterns. The and test were gathered dataset over 47965 32 characters language categorized into 9 different classes each class are very similar other’s. Classification rate with approach about 95 percent Recognition 90 percent. results show an increment rates comparison our previous work.