Artificial intelligence and machine learning are becoming increasingly important in science and society. In image processing, they are mainly used for object classification. The aim of this paper is the comparison of classical supervised machine learning methods with innovative deep learning (DL) approaches in terms of performance, which is described by the calculated accuracy. Classifiers of different characteristics are used. These are the Support Vector Machines, Random Forest, k-Nearest-Neighbor, and Naive Bayes. They are compared to two not pre-trained and four pre-trained neural networks. The former neural network are based on LeNet, the second ones include AlexNet, GoogleNet and ResNet provided by Matlab as well as a pre-trained neural network provided by MVTec HALCON. Comparisons were made using the recognition rates achieved with five real data sets from industrial applications. The results showed that not pre-trained neural networks produce worse results than classical classifiers with the given small amounts of data for training. On the other hand, the pre-trained networks achieved surpassing recognition rates. However, if there are features that describe the classes very well, the recognition performance of classical machine learning methods little differs from that of deep learning algorithms.