Analysis of Performance of KNN Classifier in Recognition of Handwritten Digits
Keywords:
Pattern Recognition, Machine learning techniques, Character Recognition, KNN, Digital Image ProcessingAbstract
The recognition of digits and manuscripts is in growing need for using in different situations like,
recognizing the handwritten postal address digits , to automatically redirect the letters in the mail and to acknowledge
the nominal values in the bank cheques. The handwritten digit recognition often faces huge difficulty when it deals
with intra-class variation because of many styles of writing, different inclination angles of the characters. Optical
Character Recognition (OCR) is a technique that is a widespread functionality in mobile devices and scanners among
others. It is used to identify and recognize the printed characters with the help of images. This paper explains the use
of the KNN (K Nearest Neighbor) algorithm used in recognition of handwritten digits. According to the results
presented, it is seen that the detection and the recognition of characters is performed with greater accuracy using the
KNN classifier and the performance is analysed.
