Issue No. 12 - Dec. (2012 vol. 34)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TPAMI.2012.40
G. Tzimiropoulos , Sch. of Comput. Sci., Univ. of Lincoln, Lincoln, UK
S. Zafeiriou , Dept. of Comput., Imperial Coll. London, London, UK
M. Pantic , Dept. of Comput., Imperial Coll. London, London, UK
We introduce the notion of subspace learning from image gradient orientations for appearance-based object recognition. As image data are typically noisy and noise is substantially different from Gaussian, traditional subspace learning from pixel intensities very often fails to estimate reliably the low-dimensional subspace of a given data population. We show that replacing pixel intensities with gradient orientations and the ℓ2 norm with a cosine-based distance measure offers, to some extend, a remedy to this problem. Within this framework, which we coin Image Gradient Orientations (IGO) subspace learning, we first formulate and study the properties of Principal Component Analysis of image gradient orientations (IGO-PCA). We then show its connection to previously proposed robust PCA techniques both theoretically and experimentally. Finally, we derive a number of other popular subspace learning techniques, namely, Linear Discriminant Analysis (LDA), Locally Linear Embedding (LLE), and Laplacian Eigenmaps (LE). Experimental results show that our algorithms significantly outperform popular methods such as Gabor features and Local Binary Patterns and achieve state-of-the-art performance for difficult problems such as illumination and occlusion-robust face recognition. In addition to this, the proposed IGO-methods require the eigendecomposition of simple covariance matrices and are as computationally efficient as their corresponding ℓ2 norm intensity-based counterparts. Matlab code for the methods presented in this paper can be found at http://ibug.doc.ic.ac.uk/resources.
principal component analysis, covariance matrices, eigenvalues and eigenfunctions, face recognition, feature extraction, gradient methods, learning (artificial intelligence), object recognition, Matlab code, subspace learning, image gradient orientations, appearance-based object recognition, pixel intensities, data population, cosine-based distance measure, principal component analysis, IGO-PCA, linear discriminant analysis, LDA, locally linear embedding, LLE, Laplacian eigenmaps, LE, Gabor features, local binary patterns, occlusion-robust face recognition, covariance matrices, eigendecomposition, ℓ2 norm intensity-based counterparts, Correlation, Principal component analysis, Robustness, Generators, Learning systems, Face recognition, Nonlinear systems, face recognition, Image gradient orientations, robust principal component analysis, discriminant analysis, nonlinear dimensionality reduction
G. Tzimiropoulos, S. Zafeiriou and M. Pantic, "Subspace Learning from Image Gradient Orientations," in IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 34, no. , pp. 2454-2466, 2012.