Handwritten Character Recognition Using Machi...

Handwritten Character Recognition Using Machine Learning and Deep Learning

Author Name : Makinedi Vasavi Devi, Marty Harshitha, Kattamuri Sai Mani Swetha, Koduri Aparna

                                                           ABSTRACT

In today’s world Machine Learning (ML) is the new Electricity. There are many fields in which both machine learning and deep learning are being used. Handwriting Recognition using these computational technologies is one of the active areas of research. It can be used in a variety of fields like Banking, Health care and many such areas where handwritten documents are dealt with.

In this project we will be working on an offline handwriting recognition system. In offline handwriting recognition systems, the handwriting of the user is available as an image. Initially we used different machine learning algorithms to perform handwritten digits recognition. Here we used KNN, PCA and SVM algorithms and neural networks. Different algorithms bring many differences in the computational speed and accuracy of recognition. In this work we achieved 96.74% for KNN, 97.77% for KNN using PCA, 97.78% for SVM and 98.19% for CNN using Relu, Relu, batch normalization dropout. For character recognition the most efficient algorithm should be implemented. Hence deep learning algorithms like CNN will help us in implementing the better solution than the others implemented here for a required task.