Issue No. 11 - Nov. (2012 vol. 34)
Yigang Peng , Dept. of Autom., Tsinghua Univ., Beijing, China
A. Ganesh , Dept. of Electr. & Comput. Eng., Univ. of Illinois at Urbana-Champaign, Urbana, IL, USA
J. Wright , Dept. of Electr. Eng., Columbia Univ., New York, NY, USA
Wenli Xu , Dept. of Autom., Tsinghua Univ., Beijing, China
Yi Ma , Dept. of Electr. & Comput. Eng., Univ. of Illinois at Urbana-Champaign, Urbana, IL, USA
This paper studies the problem of simultaneously aligning a batch of linearly correlated images despite gross corruption (such as occlusion). Our method seeks an optimal set of image domain transformations such that the matrix of transformed images can be decomposed as the sum of a sparse matrix of errors and a low-rank matrix of recovered aligned images. We reduce this extremely challenging optimization problem to a sequence of convex programs that minimize the sum of ℓ1-norm and nuclear norm of the two component matrices, which can be efficiently solved by scalable convex optimization techniques. We verify the efficacy of the proposed robust alignment algorithm with extensive experiments on both controlled and uncontrolled real data, demonstrating higher accuracy and efficiency than existing methods over a wide range of realistic misalignments and corruptions.
Robustness, Minimization, Algorithm design and analysis, Lighting, Optimization, Sparse matrices, Educational institutions, occlusion and corruption, Batch image alignment, low-rank matrix, sparse errors, robust principal component analysis
J. Wright, Wenli Xu, Yi Ma, A. Ganesh and Yigang Peng, "RASL: Robust Alignment by Sparse and Low-Rank Decomposition for Linearly Correlated Images," in IEEE Transactions on Pattern Analysis & Machine Intelligence, vol. 34, no. , pp. 2233-2246, 2012.