Image classification in cultural heritage

In this paper, an automated supervised image classification technique, specifically for classifying images in the cultural heritage domain, is developed. The developed technique classifies images according to a particular date, culture, people and historical age. The proposed technique consists of two stages, feature extraction using the unsupervised segmentation technique, and the classification stage using supervised classification techniques. Common features are extracted, and their histograms are applied to three classifiers: k-nearest neighbor (KNN), logistic regression (LR), and decision tree (DT). When our technique was applied to a repository of images from cultural heritage, it showed reduced complexity and improved classification accuracy. DT has achieved a higher weighted average recall. This is also represented by the weighted average f-measure where DT has obtained 0.81. DT has outperformed the other classifiers in terms of classifying heritage images.