An Unsupervised Content-Based Image Recognition Technique

Abstract

An unsupervised automatic moment-based image recognition technique is provided in this paper. The problem to be solved consists of classifying the images from a set, using the content similarity. In the feature extraction stage, we compute a set of feature vectors using discrete area moments. An automatic unsupervised feature vector classification approach is proposed next. It uses a hierarchical agglomerative clustering algorithm, the optimal number of clusters being determined using some validation indexes. Some experiments performed with the proposed approach are also described in this article.