Elsevier

Neurocomputing

Volume 73, Issues 10–12, June 2010, Pages 2217-2224
Neurocomputing

Robust face recognition based on illumination invariant in nonsubsampled contourlet transform domain

https://doi.org/10.1016/j.neucom.2010.01.012Get rights and content

Abstract

In order to alleviate the effect of illumination variations on face recognition, a novel face recognition algorithm based on illumination invariant in nonsubsampled contourlet transform (NSCT) domain is proposed. The algorithm first performs logarithm transform on original face images under various illumination conditions, which changes multiplicative illumination model into an additive one. Then NSCT is used to decompose the logarithm transformed images. After that, adaptive NormalShrink is applied to each directional subband of NSCT for illumination invariant extraction. Experimental results on the Yale B, the extended Yale and the CMU PIE face databases show that the proposed algorithm can effectively alleviate the effect of illumination on face recognition.

Introduction

Face recognition has become a very active research field in pattern recognition and computer vision due to its wide applications in human computer interaction, security, law enforcement and entertainment [1], [2]. It has been proved that illumination variations are more significant than the inherent differences between individuals for face recognition [3], [4]. Although various methods for face recognition have been proposed, such as PCA [5], LDA [6], LFA [7], EBGM [8], Probabilistic and Bayesian Matching [9] and SVM [10], the performance of most existing algorithms is highly sensitive to illumination variations. In order to solve the problem, a number of methods have been proposed. They fall into three groups. The first is to preprocess face images by using some techniques normalizing the images, such as logarithm transform and histogram equalization (HE), which is robust under different illumination conditions and often used for illumination normalization [11], [21]. However, it is still difficult to deal with complicated illumination variations by the above global processing techniques. Lately, block-based histogram equalization (BHE) and adaptive histogram equalization (AHE) were proposed to cope with illumination variations [24], [25]. Their performances are still not satisfactory, even though the recognition rates can be improved a little compared with HE. The second is to construct a generative 3-D face model for rendering face images with different poses and varying illumination conditions [13], [14]. A generative appearance-based model was presented for face recognition under variations in illumination conditions [13]. Its main idea is that face images under various illumination conditions can be represented by using an illumination convex cone, and the corresponding cone can be approximated in a low-dimensional linear subspace. Basri and Jacobs [14] approximated the set of convex Lambertian images under a variety of illumination conditions by a 9D linear subspace. The drawbacks of the 3D model-based methods are that many training samples under varying illumination conditions are needed and there is an assumption that the human face is a convex object, which makes this method inapplicable for practical applications. The third is to deal with illumination variations based on the Lambertian model, such as Multiscale Retinex (MSR) [28], Self Quotient Image (SQI) [15], [16], logarithmic total variation (LTV) [17], Gross and Brajovie (GB) [29], wavelet-based illumination normalization (WIN) [26], wavelet-based illumination invariant preprocessing (WIIP) [27], Multiscale facial structure representation (MFSR) [32] and illumination normalization by a series of processing (INPS) [30]. MSR deals with illumination variations by the difference between an original image and its smoothed version in logarithm domain, which is obtained by combining several low-pass filters with different cut-off frequencies. But the halo effect of MSR is serious. GB introduced by Gross and Brajovie can reduce halo effect to some extent by using an anisotropic filter. In SQL model, the illumination effect is normalized by division over a smoothed version of the image itself. This model is very simple and can be applied to a single image, but its use of the weighted Gaussian filter makes it difficult to keep sharp edges in low frequency illumination fields. LTV improves SQI by using logarithmic total variation, which can only process images under certain scale and has quite high computational expense. WIN, WIIP and MFSR attempt to normalize varying illumination by modifying wavelet coefficients. But illumination effect cannot be completely removed and Gibbs phenomena are serious. INPS is proposed to cope with illumination effect by a combination of gamma correction, difference of Gaussian filtering and contrast equalization in [30]. However, its parameter selection is usually empirical and complicated, and the number of parameters is no less than five.

To cope with the illumination variations in face recognition, we proposed a novel method employing NormalShrink filter in NSCT domain to extract illumination invariant. Compared with the existing methods, our method has the following advantages: (1) it can better preserve edges due to the nonsubsampled contourlet transform with multiscale, multidirection analysis and shift-invariance, (2) it can directly detect multiscale contour structure that is illumination invariant in the logarithm domain of a single face image, (3) the prior information (i.e., light sources assumption and large training samples) is necessary for 3D face model, but it is unnecessary for our method. Experimental results on the Yale B, the extended Yale and the CMU PIE face databases show that the proposed method is robust and effective for face recognition with varying illumination conditions.

The rest of this paper is organized as follows: Section 2 describes the NSCT-based method for illumination invariant extraction. The experimental results are presented in Section 3. Finally, we give conclusions and future work in Section 4.

Section snippets

Illumination model and logarithm transform

On the basis of the Lambertian model, a facial gray image F under illumination conditions is generally regarded as the following model [18]:F(x,y)=I(x,y)·R(x,y),where I(x,y) and R(x,y) are respectively the illumination and the reflectance at a point (x,y). R can also be regarded as an illumination invariant feature. As for robust face recognition under various illumination conditions, take R as the key facial feature due to its stability. However, it is an ill-posed problem to extract key

Experimental results and discussions

To evaluate the performance of the proposed method for illumination invariant extraction, we have applied it to three well-known databases (i.e., the Yale B, the extended Yale and the CMU PIE face databases), which are often utilized to examine the system performance when facial illumination varies. The results yielded by our method were compared with those obtained by Log [11], MSR [28], LTV [17], GB [29], WIIP [27] and INPS [30]. Quality of results is quantitatively evaluated by recognition

Conclusions and future work

In this paper, a novel NSCT-based illumination invariant extraction method is developed. The proposed method can extract illumination invariant from multiscale space. Compared with other methods, it can be directly applied in a single face image without any preprocessing and conditions. The new method can better extract geometric structure (i.e., illumination invariant) without pseudo-Gibbs phenomena around singularities and halo artifacts, which attributes to the properties of nonsubsampled

Acknowledgments

This work is partially supported by the National Science Foundation of China (Grant nos. 90820306, 60632050 and 60503026), Nanjing Institute of Technology Internal Fund (Grant no. KXJ06037) and Technology Project of provincial university of Fujian Province (Grant no. 2008F5045). Finally, the authors would like to express their heartfelt thanks to the anonymous reviewers for their constructive advice.

Yong Cheng received the M.S. degree in computer science and technology from Nanjing University of Information Science and Technology in 2004. Now, he is a Ph.D. Candidate in Nanjing University of Science and Technology, China. His research interests include image processing and pattern recognition.

References (36)

  • P. Oenev et al.

    Local feature analysis: a general statistical theory for object representation

    Network. Comput. Neural Syst.

    (1996)
  • L. Wiskott et al.

    Face recognition by elastic bunch graph matching

    IEEE Trans. Pattern Anal. Mach. Intell.

    (1997)
  • G. Guo, S.Z. Li, K. Chan, Face recognition by support vector machines, in: Proceedings of the IEEE International...
  • M. Savvides, V. Kumar, Illumination normalization using logarithm transforms for face authentication, in: Proceedings...
  • M.N. Do et al.

    The Contourlet transform: an efficient directional multiresolution image representation

    IEEE Trans. Image Process.

    (2005)
  • A.S. Georghiades et al.

    From few to many: illumination cone models for face recognition under variable illumination and pose

    IEEE Trans. Pattern Anal. Mach. Intell.

    (2001)
  • R. Basri et al.

    Lambertian reflectance and linear subspaces

    IEEE Trans. Pattern Anal. Mach. Intell.

    (2003)
  • S. Amnon et al.

    The quotient image: class-based re-rendering and recognition with varying illuminations

    IEEE Trans. Pattern Anal. Mach. Intell.

    (2001)
  • Cited by (47)

    • New set of multi-channel orthogonal moments for color image representation and recognition

      2019, Pattern Recognition
      Citation Excerpt :

      Change of intensity is the photometric transformation while the changes in orientation (Rotation), size (Scaling) and position (Translation) are geometric transformations. RST invariants of color features are more attractive than photometric invariant features [21–23] where these features achieve the invariance to both photometric and geometric transformations. Quaternion orthogonal moments were derived to compute the moments of the color images using the basic concepts of quaternion algebra [24] and the RGB color model [25] to avoid the major problems associated with the classical orthogonal moments for monochrome and gray-scale images.

    • Structured regularized robust coding for face recognition

      2016, Neurocomputing
      Citation Excerpt :

      LBP [7] and its extensions [28–31] were used to extract variation-robust texture features. To cope with the illumination variations, Cheng et al. [32] presented a robust face recognition method based on illumination invariant feature in nonsubsampled contourlet transform domain. Wu et al. [33] generalized the Weber-face to multi-scale versions for illumination-robust face recognition.

    • Joint representation and pattern learning for robust face recognition

      2015, Neurocomputing
      Citation Excerpt :

      As one of the most visible and challenging problems in computer vision and pattern recognition, face recognition (FR) has been extensively studied in the past two decades [5,25,27], and many representative methods, such as Eigenfaces [29], Fisherfaces [6], LBP [7,19], have been proposed. Moreover, to deal with challenges existed in practical FR system, active shape model [30] and active appearance model [18,31] were developed for face alignment; LBP [19] and its variants [32–35] were used to extract variation-invariant structural information; for challenging lighting changes, Cheng et al. [17] proposed a robust face recognition method based on illumination invariant feature in nonsubsampled contourlet transform domain; Wu et al. [28] presented a generalized Weber-face for illumination-robust face recognition; Hsieh and Tung [20] proposed to overcome illumination changes via shadow compensation based on facial symmetry. In order to deal with facial occlusion, Eigenimages [8–9], probabilistic local approaches [10], Markov random fields [37] and level set approach [36] were proposed.

    • Face recognition based on pixel-level and feature-level fusion of the top-level's wavelet sub-bands

      2015, Information Fusion
      Citation Excerpt :

      Various face recognition algorithms have been devised in the literature. However, up to now, face recognition is still faced with a number of challenges such as varying illumination, facial expression and poses [2–4]. Feature extraction and classification are two key steps of a face recognition system.

    View all citing articles on Scopus

    Yong Cheng received the M.S. degree in computer science and technology from Nanjing University of Information Science and Technology in 2004. Now, he is a Ph.D. Candidate in Nanjing University of Science and Technology, China. His research interests include image processing and pattern recognition.

    Yingkun Hou received the M.S. degree in applied mathematics from Shandong University of Science and Technology in 2006. He is a Ph.D. Candidate now in Nanjing University of Science and Technology, China. His research interests include image processing and pattern recognition.

    Chunxia Zhao received the B.S. degree in electric engineering and automation from Harbin Institute of Technology in 1985. She received M.S. degree in pattern recognition and artificial intelligence and Ph.D. from Harbin Institute of Technology in 1988 and 1998 respectively. Currently, she is a professor in the school of computer science and technology of Nanjing University of Science and Technology, China. Her current research interests are in the areas of pattern recognition, robot vision, image processing and artificial intelligence.

    Zuoyong Li received the B.S. degree in computer science and technology from Fuzhou University in 2002. He got his M.S. degree in computer science and technology from Fuzhou University in 2006. He is a Ph.D. Candidate now in Nanjing University of Science and Technology, China. His research interests include image segmentation and pattern recognition.

    Yong Hu is currently a Ph.D. candidate in the Department of Computer Science and Technology at Nanjing University of Science and Technology (NUST). His research interests include image processing, computer vision, and pattern recognition.

    Cailing Wang received her M.S. degree in thermal and dynamic engineering from Nanjing University of Science and Technology. She is a Ph.D. Candidate in Nanjing University of Science and Technology, China. Her research interests include image processing and pattern recognition.

    View full text