Recognition of Handwritten Devnagari Numerals

要旨

Natural language processing is a field of science and linguistics concerned with the interaction between computers and human languages. Natural language generation systems convert information from computer databases into readable human language. The term “natural ” language refers to the languages that people speak, like English and Japanese and Hindi, as opposed to artificial languages like programming languages or logic. “Natural Language processing”, programs that deal with natural language in some way or another. Character identification is one of the important subjects in the field of document Analysis and detection. Character identification can be performed on printed text or handwritten text. Printed text can be from good quality documents or degraded documents. The performance of any OCR system heavily depends upon printing quality of the input document. Little reported work has been bringing into being on the detection of degraded Devnagari Numerals. In this paper, we have predictable consider already isolated handwritten devnagari numerals on which we apply Binirazation techniques. This work is performed over 10 Devnagari numerals only. we have used structural and statistical features like Zoning, Transition features, Distance Profile features and Neighbor pixel zone etc. for generating feature sets that are used for recognizing printed Devnagari numerals by using K-NN classifiers and Parameters used for testing have achieved maximum accuracy of 90 % approximate, Squared Correlation Coefficient to get out results with 0.82 approximate with combined (Grad+ Sobel’s +Laplacian) feature vector using KNN.

主題