A Study of Structural Feature Extraction of Handwritten Numerals

Feature extraction is an important step of pattern recognition. Proposed system uses profile based feature extraction and uses Simple Profile (with cropping and without cropping image samples), Contour based feature extraction for recognizing handwritten numerals. The features are computed by using 28 X 28 as a feature length classifier used is Linear Discreminant Analysis. The classifier were trained and tested by using the MNIST Handwritten Numeral database. The average recognition rate of proposed system is observed as 87.10 % in case of simple profile without cropping the samples and also we found skeleton based feature extraction reduce the recognition result.