Handwritten English Character Recognition Using Lvq And Knn
Rasika R. Janrao · Zenodo (CERN European Organization for Nuclear Research) · 2016
A Handwritten character recognition (HCR) is an important task of detecting and recognizing in characters from the input digital image and convert it to other equivalent machine editable form. It gives high growth in image processing and pattern recognition. It has big challenges in data interpretation from language identification, bank cheques and conversion of any handwritten document into structural text form. Handwritten character recognition system uses a soft computing method like neural network, having area of research for long time with multiple theories and developed algorithm. Feature Extraction done in character recognition by introducing a new approach, diagonal based feature extraction. We used two Dataset, first one is own database of 26 alphabets, 10 numbers and 5 special characters written by various people and second is standard CEDAR database. The character recognition is carried out by supervised KNN classifier and LVQ. The results show that KNN has better results than LVQ.