Handwritten Digit Recognition from Image Using Confusion Matrix

Abstract

Recently, the creation of numerous machine learning, deep learning and computer vision algorithms has rendered hand- written visual identification by researchers highly important. I compare the results of some of the most used algorithms such as SVM , KNN & RFC, and CNN. In this article. CNN using Theano and Tensorflow for Keras. With these, I was able to reach 99.6% precision with CNN (Keras+Theano), opposed to 98.00% with SVM, 98.22% with KNN, and 97.77% with RFC.