Unsupervised SIFT-based Face Recognition Using an Automatic Hierarchical Agglomerative Clustering Solution
Tudor Barbu · Procedia Computer Science · 2013
In this paper, we propose a robust automatic unsupervised face recognition system using SIFT characteristics. A SIFT- based feature extraction is performed on the analyzed face images. Then, we introduce a novel metric for the obtained feature vectors. Next, we develop an automatic facial feature vector classification technique based on a hierarchical agglomerative clustering algorithm and some validation indexes. The recognition system described here works for large sets of faces and can be successfully applied in the face database indexing domain.