A novel adaptive illumination normalization approach is proposed to eliminate the effects caused
by illumination variations for face recognition. The proposed method divides an image into blocks and performs
discrete cosine transform (DCT) in blocks independently in the logarithm domain. For each block-DCT coefficient
except the direct current (DC) component, we take the illumination as main signal and take the reflectance
as “noise”. A data-driven and adaptive soft-thresholding denoising technique is employed in each block-DCT
coefficient except the DC component. Illumination is estimated by applying the inverse DCT in the block-DCT
coefficients, and the indirectly obtained reflectance can be used in further recognition task. Experimental results
show that the proposed approach outperforms other existing methods. Moreover, the proposed method does not
need any prior information, and none of the parameters can be determined by experience.
LIAN Zhichaoa,b* (练智超), SONG Jiea (宋 杰), LI Yanga (李 杨)
. Adaptive Illumination Normalization Approach Based on Denoising Technique for Face Recognition[J]. Journal of Shanghai Jiaotong University(Science), 2017
, 22(1)
: 45
-049
.
DOI: 10.1007/s12204-017-1797-5
[1] BARSI R, JACOBS D W. Lambertian reflectance andlinear subspaces [J]. IEEE Transactions on PatternAnalysis and Machine Intelligence, 2003, 25(2): 218-233.
[2] LEE K C, HO J, KRIEGMAN D. Acquiring linear subspacesfor face recognition under variable lighting [J].IEEE Transactions on Pattern Analysis and MachineIntelligence, 2005, 27(5): 1-15.
[3] ADINI Y, MOSES Y, ULLMAN S. Face recognition:The problem of compensating for changes in illuminationdirection [J]. IEEE Transactions on Pattern Analysisand Machine Intelligence, 1997, 19(7): 721-732.
[4] AHONEN T, HADID A, PIETIK¨AINEN M. Facedescription with local binary patterns: Applicationto face recognition [J]. IEEE Transactions on PatternAnalysis and Machine Intelligence, 2006, 28(12):2037-2041.
[5] TAN X Y, TRIGGS B. Enhanced local texture featuresets for face recognition under difficult lighting conditions[J]. IEEE Transactions on Image Processing,2012, 19(6): 1635-1650.
[6] CHEN W L, ER M J, WU S Q. Illumination compensationand normalization for robust face recognitionusing discrete cosine transform in logarithm domain[J]. IEEE Transactions on Systems, Man, and Cybernetics,Part B: Cybernetics, 2006, 36(2): 458-466.
[7] EKENEL H K, STIEFELHAGEN R. Automatic frequencyband selection for illumination robust facerecognition [C]//Proceedings of 2010 InternationalConference on Pattern Recognition. Istanbul, Turkey:IEEE, 2010: 2684-2687.
[8] CHEN T, YIN W, ZHOU X S, et al. Total variationmodels for variable lighting face recognition [J]. IEEETransactions on Pattern Analysis and Machine Intelligence,2006, 28(9): 1519-1524.
[9] M¨ULLER F. Distribution shape of two-dimensionalDCT coefficients of natural images [J]. Electronics Letters,1993, 29(22): 1935-1936.
[10] LAM E Y, GOODMAN J W. A mathematical analysisof the DCT coefficient distributions for images [J].IEEE Transactions on Image Processing, 2000, 9(10):1661-1666.
[11] DONOHO D L. De-noising by soft-thresholding [J].IEEE Transactions on Information Theory, 1995,41(3): 613-627.
[12] CHANG S G, YU B, VETTERLI M. Adaptive waveletthresholding for image denoising and compression [J].IEEE Transactions on Image Processing, 2000, 9(9):1532-1546.
[13] LIAN Z C, ER M J, CONG Y. Local line derivativepattern for face recognition [C]//The InternationalConference on Image Processing (ICIP 2012).Orlando, USA: IEEE, 2012: 1449-1452.
[14] SIM T, BAKER S, BSAT M. The CMU pose, illumination,and expression (PIE) database [C]//Proceedingsof the 5th International Conference on Automatic Faceand Gesture Recognition.Washington DC, USA: IEEE,2002: 46-51.
