In order to alleviate the influence of illumination variations on face recognition,a novel method is proposed for extracting illumination invariant from face images.In this work,the logarithm of an original face image is first decomposed by using nonsubsampled contourlet transform(NSCT),which is a fully shift-invariant,multi-scale and multi-direction transform.Then each high-pass sub-band in NSCT domain is filtered by adaptive thresholds designed according to the probability distribution of the coefficients in the sub-band.Finally,the illumination invariant is obtained by using reverse NSCT.Experimental results on the extended Yale B and the CMU PIE face data sets show that the proposed method achieves satisfactory recognition rates under varying illumination conditions.