CNN is Preferred over Other Methods for Handwritten Digit Recognition System

Handwritten digit recognition is an important problem in today’s world scenario because there are millions and millions of people across the globe with millions of different handwriting styles that could be a trouble for recognition by a human being because there are digits that are quite confusing. This paper presents an approach to offline handwritten digit recognition based on CNN and why It is preferred over other methods. For this we are using the mnist data sets for training our data and testing the outcomes to obtain the desired result to settle to the final decision that CNN is a better method than others. The only problem lies in developing an efficient algorithm for recognizing handwritten digits mainly the ones that are a little bit confusing. This method can be very useful for the future like the monitoring of the vehicles that break the traffic rules, and in banks for detecting fraud cheques.