18-10-2014, 09:46 AM
Abstracts: Handwritten character recognition system is used to convert documents which are handwritten into such a format that can be understood by the computer machine as well as human being. The input given is the image of handwritten text. The output generated is the text file, readable by computer as well as human beings. The handwritten character recognition system has 5 stages - getting the image for input, binarizing the image, segmenting the characters, extracting the features and recognizing the characters. The main emphasis of our project is on the feature extraction of the characters. The methods implemented for feature extraction are – Template Matching, Image Centroid based zoning, Density based zoning, Histogram Projection. Three new methods have been implemented by hybridizing the existing methods – Image Centroid based zoning along with Density based zoning, Image Centroid based zoning along with Histogram Projection, Density based Zoning along with Histogram Projection.