Open AccessBroken Character Recognition using Connected Components and Convolutional Neural Network
Publication year - 2020
Publication title -
international journal of recent technology and engineering
Language(s) - English
Resource type - Journals
ISSN - 2277-3878
DOI - 10.35940/ijrte.e1011.0285s20
Subject(s) - convolutional neural network , character (mathematics) , computer science , artificial intelligence , character recognition , component (thermodynamics) , pattern recognition (psychology) , optical character recognition , natural language processing , connected component , artificial neural network , speech recognition , image (mathematics) , mathematics , physics , geometry , thermodynamics
Recognizing broken characters in scanned and ancient scanned text document is not easy because the characters may be broken and unclear. Many researches have been carried to recognize these broken characters. In this research paper we have described a new broken characters recognition method for English text documents only. The proposed method uses a hybrid approach which uses connected component concepts and convolutional neural network to identify the broken characters. The input to the approach is scanned or ancient text document which contains unclear text that is difficult to recognize and hence our new proposed methodology will recognize these characters with greater accuracy and it will give the recognized characters to the user. The projected technique has attained a precision up to 92% in recognition.