Comparison Analysis of Traditional Machine Learning and Deep Learning Techniques for Data and Image Classification

The purpose of the study is to analyse and compare most common machine learning deep techniques used for computer vision 2D object classification tasks. Firstly, we will present theoretical background Bag Visual words model Deep Convolutional Neural Networks (DCNN). Secondly, implement a Words model, VGG16 CNN Architecture. Thirdly, our custom novice DCNN in which test aforementioned implementations on modified version Belgium Traffic Sign dataset. Our results showcase effects hyperparameters traditional advantage terms accuracy DCNNs compared classical methods. As tests indicate, proposed solution can achieve similar - some cases better than existing architectures. Finally, technical merit this article lies presented computationally simpler architecture, believe pave way towards using more efficient architectures basic