Algorithm Comparison for Cultural Heritage Image Classification

Abstract

Digitization represents an important part of the development of online systems. As such it includes, among other, the deployment, categorization and preservation of audio, video and textual contents online. Such process is especially interesting from the perspective of cultural heritage, as it allows the long-term preservation and sharing of culture worldwide. This study observes four classification algorithms: (i) themultilayer perceptron, (ii) averaged one dependence estimators, (iii) forest by penalizing attributes, and (iv) the k-nearest neighbor rough sets and analogy based reasoning, before and after attribute classifcation, and compares these with the results obtained from the convolutional neural network. The obtained results show that the best classification performance was achieved by the multilayer perceptron, followed by the convolutional neural network.