Recognition of Online Handwritten Punjabi Numerals

Abstract

The main goal of this thesis is to develop an online handwritten Gurmukhi numeral recognition system.In this,we have carried out the data collection,pre-processing and the recognition process.In pre-processing ,four algorithms: normalization, interpolation, resampling and smoothing have been implemented.The algoriths are implimented in JAVA.Recognition is done using SVM.We have used three kernel function: Linear,Polynomial and RBF of SVM for recognition and the results obtained are resonably good.The highest accuracy achieved is 97.5% when linear kernel is used.