Ensemble of patterns of oriented edge magnitudes descriptors for face recognition
In this work we propose an ensemble of descriptors for face recognition. Starting from the base patterns of the oriented edge magnitudes (POEM) descriptor, we developed different ensembles by varying the preprocessing techniques, the parameters for extracting the accumulated magnitude images (AM), and the parameters of the local binary patterns (LBP) applied to AM. Our best proposed ensemble works well regardless of whether dimensionality reduction by principal component analysis (PCA) is performed or not before the matching step. We validate our approach using the FERET datasets and the Labeled Faces in the Wild (LFW) dataset. We obtain very high performance rates in both datasets. To the best of our knowledge, we obtain one of the highest performances reported in the literature on the FERET datasets. We want to stress that our ensemble obtains these results without combining different texture descriptors and without any supervised approach or transform. Finally, two cloud use cases are proposed. The MATLAB source of our best approach will be freely available: http://www.dei.unipd.it/wdyn/?IDsezione=3314&IDgruppo_pass=124
Information Technology and Cybersecurity
Ensemble of descriptors, Face recognition, Local binary patterns, Patterns of oriented edge magnitudes
Nanni, Loris, Alessandra Lumini, Sheryl Brahnam, and Mauro Migliardi. "Ensemble of patterns of oriented edge magnitudes descriptors for face recognition." In Proceedings of the International Conference on Image Processing, Computer Vision, and Pattern Recognition (IPCV), pp. 1675-680. The Steering Committee of The World Congress in Computer Science, Computer Engineering and Applied Computing (WorldComp), 2013.
Proceedings of the 2013 International Conference on Image Processing, Computer Vision, and Pattern Recognition, IPCV 2013