Analysis on Devanagari Text Generation by Using Machine Learning Techniques

Abstract

Many Authors are developed different methods for recognition of the Devanagari script. The existing method is processing to recognize the Devanagari script which is discussed with notable performances. Generally, the recognition process mainly consists of three steps pre-processing, extraction of feature and finally classification. For character recognition different algorithm were being developed with different advances which include neural network algorithm (NNA), pattern matching algorithm (PMA), structural algorithm (SA), support vector machine algorithm (SVMA), statistical algorithm, hidden Markov model (HMM) and template matching algorithm(TMA). In template matching algorithm only the typewritten characters can be recognized. But the other algorithm like neural network algorithm (NNA), structural algorithm (SA), and support vector machine (SVM) can recognize both  handwritten and typewritten Every algorithm contains both advantages and disadvantages. The main motivation of this thesis work is to overcome the above mention the drawbacks.