Developing an Efficient CNN Model to Recognize Handwritten Digits

Abstract

The work presented in this thesis is focussed on the pattern recognition, computer vision and image classification. Among the most used methods in deep learning field DL, we find CNNs (Convolutional Neural Networks) which can be considered as the best used technique in the field. Effectively, we have developed an automatic classifier that permits to classify some given grayscale images representing handwritten digits into one of 10 classes (digits from 0 to 9), inclusively. For this purpose, we have used ML and DL approaches. First, we proceeded to the classification task using many ML algorithms including: LR, LDA, KNN, CART, NB, and SVM. Second, we proposed a new CNN model composed of many convolutional layers. We have also explained the obtained results. Finally, we have established a comparison between different algorithms.