|Table of Contents|

Illumination Invariant and Face Recognition Based on NSCT and SQI

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

Issue:
2010年04期
Page:
425-430
Research Field:
Publishing date:

Info

Title:
Illumination Invariant and Face Recognition Based on NSCT and SQI
Author(s):
CHENG Yong12ZHAO Chun-xia1WANG Huan1WU Gang12YANG Ying-yu1
1.School of Computer Science and Technology,NUST,Nanjing 210094,China;2.School of Communication Engineering,Nanjing Institute of Technology,Nanjing 211167,China
Keywords:
face recognition nonsubsampled contourlet transform self-quotient images illumination invariant
PACS:
TP391.41
DOI:
-
Abstract:
In order to eliminate the effect of varying illumination on face recognition,a novel illumination invariant method based on nonsubsampled contourlet transform(NSCT) and self-quotient image(SQI) is proposed.The method first performs Gamma correction on image under various lighting conditions,which can decrease the effect of varying illumination to some extent.The NSCT is used for multiresolution analysis.NormalShrink filtering is applied to high frequency subbands and a smooth image can be obtained by inverse nonsubsampled contourlet transform.A self-quotient image is used for illumination invariant extraction.Experimental results from Yale B and CMU PIE databases show that the proposed method can effectively eliminate the effect of varying illumination on face recognition,and the recognition rate of the method here is higher than the multiscale Retinex,self-quotient image and logarithmic total variation methods.

References:

[ 1] ?? Che llappa R, W ilson C L, S irohey S. Hum an andm ach ine recogn ition o f faces: a survey [ J]. Proceedings of the IEEE, 1995, 83( 5): 705- 740.
[ 2] Bowyer K W, Chang K, Flynn P J. A survey of approaches and challenges in 3D and m ult-i modal 3D + 2D face recogn ition[ J]. Compu ter V ision and Im ag e Unde rstanding, 2006, 101( 1): 1- 15.
[ 3] Shan Shiguang, GaoW en, Cao Bo, e t a.l Illum ina tion norm a liza tion for robust face recogn ition aga inst vary ing illum ination conditions[ A]. Proceedings o f IEEE Interna tiona l Conference on Autom a tic Face and G esture Recognition[ C ]. W ashing ton DC, USA: IEEE Computer Soc ie ty, 2003: 157- 164.
[ 4] P ize r SM, Amburn E P, Austin J D, et a .l Adaptiv e h istogram equa lization and its var iations[ J]. Compu ter V ision, G raph ics, and Im ag e Pro cessing, 1987, 39( 3): 355- 368.
[ 5] Xie Xudong, Lam K M. Face recogn ition under va rying illum ina tion based on a 2D face shapem ode l[ J]. Patte rn Recognition, 2005, 38( 2): 221- 230.
[ 6] Geo rghiades A S, Be lhum eur P N, Kr iegm an D J. From few to m any: illum ination cone m ode ls fo r face recognition under va riab le lighting and pose[ J]. IEEE Transactions on Pattern Analysis andM ach ine Inte ll-i gence, 2001, 23( 6) : 643- 660.
[ 7] Basri R, Jacobs D W. Lambertian reflectance and linear subspaces[ J]. IEEE Transactions on Pattern Analysis andMach ine Intelligence, 2001, 25( 2): 218- 233.
[ 8] Jobson D J, Rahm an Z, W oode ll G A. A m ultisca le Retinex fo r br idg ing the gap betw een co lo r images and the hum an observation o f scenes [ J]. IEEE Transactions on Im age Process ing, 1997, 6( 7): 965- 976.
[ 9] W angH a itao, L i S Z, W ang Yangsheng. Face recogn ition under vary ing lighting cond itions using se lf quotient im ag e [ A ] . Proceed ing s of IEEE Internationa l Conference on Au tom atic Face and Gesture Recogn-i tion [ C ]. Seou ,l Ko rea: IEEE Computer So ciety, 2004: 819- 824.
[ 10] 王海涛, 刘俊, 王阳生. 自商图像[ J]. 计算机工程, 2005, 31( 18) : 178- 179.
[ 11] Chen T, Y inWo tao, Zhou X iangsean, et a.l Tota l var-i ation m ode ls for var iable lighting face recogn ition [ J]. IEEE Transactions on Pa ttern Ana lysis andM ach ine Inte lligence, 2006, 28( 9): 1519- 1524.
[ 12] Gross R, Bra jov icV. An im ag e preprocessing algo rithm for illum ination invariant face recogn ition[ A ]. The 4th Inte rna tiona l Conference on Audio-and V ideo-Based B-i om etr ic Person Authentication ( AVBPA) [ C ]. Gu ildford, UK: Spr inge r, 2003: 10- 18.
[ 13] H orn B K P. Robot V is ion[M ]. Cam bridgeMA: M IT Press, 1986.
[ 14] Do M N, VetterliM. The contourlet transform: an eff-i cient directiona lm ultiresolution mi age representation[ J]. IEEE Transactions on Im age Processing, 2005, 14( 12): 2091- 2106.
[ 15] CunhaA L, Zhou J ianp ing, DoM N. The nonsubsampled contour let transform: theory, design, and appl-i cations[ J] . IEEE Transactions on Im age Processing, 2006, 15( 10): 3089- 3101.
[ 16] K aur L, Gupta S, Chauhan R C. Im ag e deno ising using wave let thresho lding [ A ]. Proceed ing s o f the Th ird Indian Conference on Compu ter V ision, Graphics and Im age Processing [ C ]. Ahm adabad: A llied Publishers Priva te L im ited, 2002: 1522- 1531.
[ 17] H uang L iang liang, W ang H ongzh,i Zhu Bo. Adaptiv e thresho lds a lgo rithm o f im age deno ising based on nonsubsamp led contourlet transfo rm [ A ]. Proceedings o f the IEEE Internationa l Conference on Com puter Sc-i ence and Softw are Eng inee ring[ C ]. W ash ing ton DC, USA: IEEE Com puter So ciety, 2008: 209- 212.
[ 18] Sim T, Bake r S, BsatM. The CMU Pose, Illum ination, and Expression Database[ J] . IEEE Transactions on Patte rn Analysis and M ach ine Inte lligence, 2003, 25( 12): 1615- 1618.

Memo

Memo:
-
Last Update: 2012-11-02