Comparative Study of Handwritten Character Recognition System for Indian Languages

Abstract

AbstractHandwritten Character Recognition is the emerging domain for the research in image processing and pattern recognition because a huge amount of handwritten and OCR documents need to be processed in daily needs of different systems. Based on surveys of Human Development very less amount of Indian population can write and read English language. It specifies the need for research in Indian script for OCR technology. The research in this area is actually needed for developing countries such as India and remains beneficial to inspire the human-machine interface system. HCR makes it ease the automatic processing of tax and admission forms, bank cheques, addresses and postal codes. In the comparison of printed character recognition, handwritten character recognition is difficult because it can be different from person to person based on the style of writing, curve, size, strokes and thickness of character. Sufficient research work has been performed in foreign handwritten characters such as Arabic, Chinese and Japanese but very less research work has been performed in Indian scripts. This paper provides a comparative study of famous and well-known Tamil, Telugu, Kannada and Gujarati Indian scripts. This paper provides a detailed study based on the used methodology, dataset and accuracy.KeywordsHandwritten Text RecognitionCharacter RecognitionHandwritten Character RecognitionHCROptical Character RecognitionOCRIndian Language Text RecognitionTelugu Character RecognitionTamil Character RecognitionKannada Character RecognitionGujarati Character Recognition