Multilinear Sparse Decomposition for Best Spectral Bands Selection
| Affiliation auteurs | !!!! Error affiliation !!!! |
| Titre | Multilinear Sparse Decomposition for Best Spectral Bands Selection |
| Type de publication | Conference Paper |
| Year of Publication | 2014 |
| Auteurs | Bouchech HJamel, Foufou S, Abidi M |
| Editor | Elmoataz A, Lezoray O, Nouboud F, Mammass D |
| Conference Name | IMAGE AND SIGNAL PROCESSING, ICISP 2014 |
| Publisher | European Assoc Image & Signal Proc; Int Assoc Pattern Recognit; Inst Univ Technologie; Univ Caen Basse Normandie; ENSICAEN; Ctr Natl Rech Sci; Conseil Reg Basse Normandie; Conseil Gen Manche; Communaut Urbaine Cherbourg |
| Conference Location | HEIDELBERGER PLATZ 3, D-14197 BERLIN, GERMANY |
| ISBN Number | 978-3-319-07998-1; 978-3-319-07997-4 |
| Mots-clés | HGPP, MBLBP, Multilinear, sparse, Spectral bands, Tensor |
| Résumé | Optimal spectral bands selection is a primordial step in multispectral images based systems for face recognition. In this context, we select the best spectral bands using a multilinear sparse decomposition based approach. Multispectral images of 35 subjects presenting 25 different lengths from 480nm to 720nm and three lighting conditions: fluorescent, Halogen and Sun light are groupped in a 3-mode face tensor T of size 35x25x2. T is then decomposed using 3-mode SVD where three mode matrices for subjects, spectral bands and illuminations are sparsely determined. The 25x25 spectral bands mode matrix defines a sparse vector for each spectral band. Spectral bands having the sparse vectors with the lowest variation with illumination are selected as the best spectral bands. Experiments on two state-of-the-art algorithms, MBLBP and HGPP, showed the effectiveness of our approach for best spectral bands selection. |