Texture based features for robust palmprint recognition: a comparative study
 3.9k Downloads
 8 Citations
Abstract
Palmprint is a widely used biometric trait deployed in various accesscontrol applications due to its convenience in use, reliability, and low cost. In this paper, we propose a novel scheme for palmprint recognition using a sparse representation of features obtained from Bank of Binarized Statistical Image Features (BBSIF). The palmprint image is characterized by a rich set of features including principal lines, ridges, and wrinkles. Thus, the use of an appropriate texture descriptor scheme is expected to capture this information accurately. To this extent, we explore the idea of BBSIF that comprises of 56 different BSIF filters whose responses on the given palmprint image is processed independently and classified using sparse representation classifier (SRC). Extensive experiments are carried out on three different largescale publicly available palmprint databases. We then present an extensive analysis by comparing the proposed scheme with seven different contemporary stateoftheart schemes that reveals the efficacy of the proposed scheme for robust palmprint recognition.
Keywords
Biometrics Palmprint Comparative study Texture features1 Introduction
Biometric systems are widely used in access control and securitybased applications. The goal of the biometric system is to utilize physical and/or behavior characteristics to identify/verify the subject of interest. There exist various kinds of biometric systems that are based on physical and/or behavioral cues such as the face, iris, speech, keystroke, palmprint, retina, and so on. Among these, the palmprintbased biometric system that has been investigated for over 15 years has demonstrated its applicability as a successful biometric modality. Palmprints exhibit a unique characteristic that can be characterized using texture features that are contributed due to the presence of palm creases, wrinkles, and ridges. Furthermore, the palmprints can be captured using lowcost sensors with a very lowresolution imaging of 75 dotsperinch (dpi) [1, 2]. Further, recent work [3] has demonstrated the antispoofing nature of palmprints that places the palmprint as a highly reliable biometric characteristic.
Characteristics of palmprint recognition approaches
Feature type  Complexity  Accuracy 

Local feature approach  High  High 
Statistical approach  Low  Low 
Appearancebased approach  Low  Medium 
Texturebased method (or) texture coding  Medium  High 
Hybrid method  High  High 

A new method based on the Bank of BSIF (BBSIF) and sparse representation classifier (SRC) for palmprint recognition.

Extensive experiments are carried out on the following three different palmprint databases, namely: PolyU contact palmprint database [36] with 356 subjects, IIT Delhi contactless palmprint database [37] with 236 subjects, and Multispectral palmprint PolyU database [3] with 500 subjects.

Comprehensive analysis by comparing the proposed scheme with seven different stateoftheart contemporary schemes based on LBP [24], palmcode [25], ordinal code [26], fusion code [27], Gabor transform with KDA [13], Gabor transform with sparse representation [38], and also with our previously proposed technique based on the sparse representation of BSIF [35].
All in all, the proposed framework, being a simple, novel, and first of its kind in the literature of palmprint verification/recognition, is expected to open up a new dimension for further research in the field of palmprint biometrics.
The rest of the paper is structured as follows: Section 2 presents the proposed scheme for robust palmprint recognition, Section 3 discusses the experimental setup, protocols, and results, and Section 4 draws the conclusion.
2 Proposed method
2.1 Region of interest extraction
The main idea of the region of interest (RoI) is to extract the significant region from the palmprint that constitutes for the rich set of features such as principal lines, ridges, and wrinkles by compensating for rotation and translation. The accurate extraction of RoI plays a crucial role in improving the performance of the overall palmprint recognition. In this work, we have employed the algorithm proposed in [23] which is based on aligning the palmprint by computing the center of mass and also by locating the valley regions. We carried out this RoI extraction scheme only on the PolyU palmprint database as the other two databases (MSPolyU and IITD) have already provided the RoI images.
2.2 Bank of BSIF features and sparse representation classifier
The idea behind the proposed BBSIF is to construct a bank of filters that is trained using a set of natural images. Traditionally, one can train BBSIF filters in an unsupervised manner using the most popular techniques, namely: Restricted Boltzmann Machines (RBMs) [39, 40], autoencoders [41], sparse coding [42], and independent component analysis [43, 44]. Among these schemes, the use of ICA is a more appealing choice as it overcomes the tuning of large sets of hyperparameters and can also provide a statistically independent basis that in turn can be used as the filter to extract the features from the given image. Thus, given the natural images, we first normalize to have a zero mean and unit variance [34]. Then, we sample N _{ Im } number of patches to learn the BSIF filters using ICA. Thus, the size of the image patch sampled from natural images will fix the size of the BSIF filter to be learned and the selection of the number of top ICA basis will indicate the length of the BSIF filter. For instance, BSIF filter with the size 5×5 and length 8 corresponds to the top 8 basis of the ICA algorithm learned using an image patch size of 5×5 sampled from the natural image. Thus, by varying both size and length, one can learn some BSIF filters from the natural images. In this work, we consider 56 different prelearned filters with varying size and length that can constitute a Bank of BSIF (BBSIF) filter.
To achieve the good performance in palmprint recognition using BSIF, we need to consider two important factors, namely: filter size and filter length. However, the use of a single filter with a fixed length may not be capable of capturing sufficient information to achieve accurate palmprint recognition. Thus, in this work, we propose to use the bank of filters with varying filter size and length. The filter size is varied from 5×5 to 17×17 in steps of two such that we have filters of seven different sizes. In a similar manner, we also vary the length of the filter (or a number of the independent components) from 5 till 12 in steps of 1 to get 8 different lengths. Thus, our ensemble has 7×8=56 filters such that the response for palmprint image is obtained independently. Given a palmprint sample P(m,n), we get 56 independent BSIF coded images R _{ P }={R _{ P1},R _{ P2},…,R _{ P56}}.
 1.Given the reference palmprint samples, we first extract the BSIF features (corresponding to one filter) and construct a training T _{ r } for all C classes (or subjects) as follows:$$ T_{r} = \left[T_{r1}, T_{r2}, \ldots,T_{rC}\right] \in \Re^{N \times \left(n_{u}.C\right)} $$(4)
Where, n _{ u } denotes the number of reference samples for each class and N indicates the dimension of the BSIF features obtained on n _{ u } reference samples from C classes (or subjects).
 2.Given that the test (or probe) sample T _{ e } obtains the BSIF features (corresponding to same filter as above) that can be considered as a linear combination of the training vectors as:$$ T_{e} = T_{r}\alpha $$(5)Where,$$ \alpha = \left[ \alpha_{1}, \ldots, \alpha_{1n_{u}},\alpha_{2}, \ldots, \alpha_{2n_{u}}, \ldots, \alpha_{C1}, \ldots,\alpha_{{Cn}_{u}}\right] $$(6)
 3.Solve l _{1} minimization problem [46] as follows:$$ \hat\alpha = \arg \min_{\alpha^{'} \in \Re^{N}} \ \alpha^{'}\_{1} T_{e} = T_{r}\alpha^{'} $$(7)
 4.Calculate the residual as follows:$$ r_{c}(y) = \T_{e}  \Pi_{C}(\alpha^{'})\_{2} $$(8)
 5.
Finally, obtain the comparison score as the residual errors to compute the performance of the overall system.
Finally, we repeat the above mentioned steps 1–5 on all 56 different BSIF filters in the bank and obtain the final comparison score that corresponds to the minimum of residual errors obtained on all 56 filters in the bank.
3 Experimental results and discussion
This section presents the experimental results obtained on the proposed scheme for palmprint recognition. Extensive experiments are carried out on three different largescale publicly available palmprint databases such as: (1) PolyU palmprint database [36], (2) IIT Delhi palmprint database [37], and (3) Multispectral palmprint PolyU database [3]. All the experimental results are presented in terms of equal error rate (EER), and we also present the statistical validation of the results with 90 % confidence interval [13]. In the following section, we present the experimental protocol adopted in this work.
3.1 Assessment protocol
This section describes the evaluation protocol adopted in this work on three different palmprint databases that are the same with our previous paper [35].
PolyU palmprint database This database comprises of 352 subjects such that each subject has ten samples collected in two different sessions. For our experiments, we consider all ten samples from the first session as a reference and all samples from the second session as probe samples. The database is available for research purposes at http://www4.comp.polyu.edu.hk/~biometrics/MultispectralPalmprint/MSP.htm.
IIT Delhi palmprint database This database consists of 235 subjects with both left and right palmprint samples. Each subject has five samples captured independently from both left and right palmprints. To evaluate this database, we consider four samples as the reference and remaining one sample as the probe sample. We repeat this selection of reference and probe samples using leavingoneout cross validation with k = 10, and finally, we present the result by averaging the performance overall ten runs. The database is available for research purposes at http://www4.comp.polyu.edu.hk/~csajaykr/IITD/Database_Palm.htm.
MultiSpectral PolyU palmprint database This database consists of 500 subjects whose palmprint samples are captured in two different sessions in four different spectra: blue, red, green, and near infrared (NIR). Each session has six samples per subject. Thus, we select samples from first session as reference samples while we select second session samples as a probe. We repeat this procedure for all four spectral bands, and results are presented independently. The database is available for research purposes at http://www4.comp.polyu.edu.hk/~biometrics/MultispectralPalmprint/MSP.htm.
3.2 Results and discussion
Performance of the proposed method on PolyU palmprint database
Methods  EER (%) with 

90 % confidence interval  
Proposed scheme  4.06 [3.56; 4.56] 
BSIFSRC [35]  6.19 [5.09; 7.29] 
LGSRC [38]  7.67 [6.47; 8.87] 
LGKDA [13]  7.96 [6.56; 9.36] 
Palm Code [25]  14.66 [13.26; 16.06] 
Ordinal Code [26]  7.66 [6.46; 8.86] 
Fusion Code [27]  14.51 [13.21; 15.81] 
LBPSRC [24]  46.22 [43.84; 48.62] 
Performance of the proposed method on IIT Delhi palmprint database
IIT Delhi DB  Methods  EER (%) 

90 % confidence interval  
Proposed scheme  0.12 [0.05;0.19]  
BSIFSRC [35]  0.42 [0.32;0.52]  
GaborSRC [38]  1.23 [0.93; 1.53]  
Ordinal code [26]  0.20 [0.15; 0.25]  
Left hand  GaborKDA [13]  2.34 [1.54; 3.14] 
Palmcode [25]  2.67 [1.97; 9.37]  
Fusion code [27]  2.34 [1.54; 3.14]  
LBPSRC [24]  10.41 [8.81;12.10]  
Proposed scheme  0.72 [0.32; 1.12]  
BSIFSRC [35]  1.31 [0.91; 1.71]  
GaborSRC [38]  1.42 [1.02; 1.82]  
Ordinal code [26]  1.89 [1.59; 2.19]  
GaborKDA [13]  7.82 [6.42; 9.22]  
Right hand  Palmcode[25]  3.41 [2.91; 3.91] 
Fusion code [27]  3.39 [2.89; 3.89]  
LBPSRC [24]  13.38 [11.49; 15.28] 
Performance of the proposed method on MS PolyU palmprint database
Spectrum  Methods  EER (%) 

90 % confidence interval  
Proposed scheme  0  
BSIFSRC [35]  0  
GaborSRC [38]  0  
Ordinal code [26]  0  
GaborKDA [13]  0.7 [0.55; 1.25]  
Blue  Palmcode [25]  0.2 [0.1; 0.3] 
Fusion code [27]  0.4 [0.28; 0.68]  
LBPSRC [24]  9.76 [8.36; 11.16]  
Proposed scheme  0  
BSIFSRC [35]  0  
GaborSRC [38]  0  
Ordinal code [26]  0  
GaborKDA [13]  0.40 [0.3; 0.5]  
Green  Palmcode [25]  0.41 [0.21; 0.61] 
Fusion code [27]  0.67 [0.38; 0.96]  
LBPSRC [24]  17.75 [15.58; 19.92]  
Proposed scheme  0  
BSIFSRC [35]  0  
GaborSRC [38]  0  
Ordinal code [26]  0  
GaborKDA [13]  0.40 [0.3; 0.5]  
Red  Palmcode [25]  0 
Fusion code [27]  0.21 [0.03; 0.38]  
LBPSRC [24]  14.78 [13.58; 15.98]  
Proposed scheme  0  
BSIFSRC [35]  0  
GaborSRC [38]  0  
Ordinal code [26]  0  
GaborKDA [13]  0.70 [0.4: 1.0]  
NIR  Palmcode [25]  0 
Fusion code [27]  0.2 [0.02;0.38]  
LBPSRC [24]  14.78 [13.38; 16.18] 
Thus, from the above experiments, it can be observed that the proposed scheme has shown the best performance when compared with five wellestablished stateoftheart schemes for the palmprint recognition. Further, the performance achieved using the proposed scheme on three different databases justifies its robustness and applicability of the palmprint recognition.
Computation time of different algorithm used in this work
Algorithms  Computation time in seconds 

Proposed scheme  30 
BSIFSRC  12 
GaborSRC  25 
Ordinal code  15 
GaborKDA  12 
Palmcode  5 
Fusion code  12 
LBPSRC  10 
4 Conclusions
Accurate representation of the features plays a vital role in improving the accuracy and reliability of the palmprint recognition. In this paper, we have introduced a novel approach for the palmprint recognition based on BBSIF and SRC. The main idea of the proposed method is to use multiple BSIF filters with various size and length to constitute an ensemble (or bank of BSIF filters). Since each of these BSIF filters are learned on the natural images using the independent component analysis (ICA), they exhibit the property of statistical independence. We proposed to build the BBSIF with 56 different BSIF filters. Then, each of these filters is associated with the SRC that essentially perform the sparse representation of each BSIF filter. Thus, given a palmprint sample, we obtain its response on each of the BSIF filter and then obtain the corresponding comparison score using SRC. Finally, we select the best comparison score that corresponds to the minimum value of the residual error. The proposed method is validated by conducting extensive experiments on three different largescale publicly available databases that indicated the outstanding performance. The performance of the proposed scheme is compared with seven wellestablished stateoftheart schemes. The obtained results justify that the proposed scheme has emerged as an efficient and robust tool for accurate palmprint recognition.
Notes
Acknowledgements
The authors extend their thanks to reviewers for the comments and suggestions which helped in improving the overall quality of this article. This work was funded by the EU 7^{ th } Framework Program (FP7) under grant agreement n ^{ o } 284862 for the largescale integrated project FIDELITY.
References
 1.D Zhang, Palmprint authentication (Springerverlag, Guangzhou, China, 2004).Google Scholar
 2.A Genovese, V Piuri, F Scotti, Advances in information security, vol. 60. Springer, (2014). doi:http://dx.doi.org/9783319103655
 3.D Zhang, Z Guo, G Lu, YLL Zhang, W Zuo, Online joint palmprint and palmvein verification. Expert Syst. Appl. 38(3), 2621–2631 (2011).CrossRefGoogle Scholar
 4.W Shu, D Zhang, in Pattern Recognition, 1998. Proceedings. Fourteenth International Conference On, 1. Palmprint verification: an implementation of biometric technology, pp. 219–221. Brisbane, Qld, 1998.Google Scholar
 5.D qingyun, Y yinglin, Z dapeng, A line feature extraction method based on morphological median pyramid. J. South China Univ. Technol. 29(5), 14–18 (2001).Google Scholar
 6.CC Han, HL Cheng, CL Lin, KC Fan, Personal authentication using palmprint features. Pattern Recognit. 36(2), 371–381 (2003).CrossRefGoogle Scholar
 7.JA Noble, Finding corners. Image Vis. Comput. 6(2), 121–128 (1988).CrossRefGoogle Scholar
 8.L Liu, D Zhang, in IEEE International Conference on Image Processing (ICIP), 3. Palmline detection, (2005), pp. 269–272. doi:http://dx.doi.org/10.1109/ICIP.2005.1530380
 9.JY Gan, DP Zhou, in Signal Processing, 2006 8th International Conference On, 3. A novel method for palmprint recognition based on wavelet transform, pp. 1–7. Beijing, 2006.Google Scholar
 10.X yuli, Palmprint feature extraction based on low frequency distribution. Microcomput. Appl. 20(1), 40–43 (2011).Google Scholar
 11.MMM Fahmy, Palmprint recognition based on mel frequency cepstral coefficients feature extraction. Ain Shams Eng. J. 1(1), 39–47 (2010).CrossRefGoogle Scholar
 12.Y Zhang, D Zhao, G Sun, Q Guo, B Fu, in Artificial Intelligence and Computational Intelligence (AICI), 2010 International Conference On, 1. Palm print recognition based on subblock energy feature extracted by real 2dgabor transform, pp. 124–128. Sanya, 2010.Google Scholar
 13.R Raghavendra, B Dorizzi, A Rao, GH Kumar, Designing efficient fusion schemes for multimodal biometric systems using face and palmprint. Pattern Recognit. 44(5), 1076–1088 (2011).CrossRefzbMATHGoogle Scholar
 14.P xin, R qiuqi, W yanxia, Palmprint recognition using gabor local relative features. Comput. Eng. Appl. 48(15), 34–38 (2012).Google Scholar
 15.GS Badrinath, N Kachhi, P Gupta, Verification system robust to occlusion using loworder zernike moments of palmprint subimages. Telecommun. Syst. 47(3–4), 275–290 (2011).CrossRefGoogle Scholar
 16.G Lu, D Zhang, K Wang, Palmprint recognition using eigenpalms features. Pattern Recognit. Lett. 24(9–10), 1463–1467 (2003).CrossRefzbMATHGoogle Scholar
 17.H Sang, W Yuan, Z Zhang, in Advances in Neural Networks – ISNN 2009. Lecture Notes in Computer Science, 5552, ed. by W Yu, H He, and N Zhang. Research of palmprint recognition based on 2dpca (WuhanChina, 2009), pp. 831–838.CrossRefGoogle Scholar
 18.W Zuo, K Wang, D Zhang, in Image Processing, 2005. ICIP 2005. IEEE International Conference On, 2. Bidirectional pca with assembled matrix distance metric, (2005), pp. 958–61. doi:http://dx.doi.org/10.1109/ICIP.2005.1530216
 19.X Pan, QQ Ruan, Palmprint recognition using Gabor featurebased (2d)2pca. Neurocomputing. 71(13–15), 3032–6 (2008).CrossRefGoogle Scholar
 20.GM Lu, KQ Wang, D Zhang, in Machine Learning and Cybernetics, 2004. Proceedings of 2004 International Conference On, 6. Wavelet based independent component analysis for palmprint identification, (2004), pp. 3547–35506. doi:http://dx.doi.org/10.1109/ICMLC.2004.1380404
 21.X Wu, D Zhang, K Wang, Fisherpalms based palmprint recognition. Pattern Recognit. Lett. 24(15), 2829–38 (2003).CrossRefGoogle Scholar
 22.M Ekinci, M Aykut, Gaborbased kernel pca for palmprint recognition. Electron. Lett. 43(20), 1077–9 (2007).CrossRefGoogle Scholar
 23.R Raghavendra, A Rao, GK Hemantha, in International Conference on Advances in Computing, Control, Telecommunication Technologies. A novel three stage process for palmprint verification (Trivandrum, Kerala, 2009), pp. 88–92.Google Scholar
 24.X Wang, H Gong, H Zhang, B Li, Z Zhuang, in Pattern Recognition, 2006. ICPR 2006. 18th International Conference On, 3. Palmprint identification using boosting local binary pattern (IEEEHong Kong, 2006), pp. 503–6.Google Scholar
 25.A Kumar, H Shen, in 3rd Int Conference on Image and Graphics, ICIG20. Palmprint identification using palm codes (Hong Kong, pp. 258–261. http://dx.doi.org/10.1007/9783642040702_42
 26.Z Sun, T Tan, Y Wang, SZ Li, in Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference On. Ordinal palmprint represention for personal identification [represention read representation], (2005), pp. 279–84. http://dx.doi.org/10.1109/CVPR.2005.267
 27.A Kong, D Zhang, M Kamel, Palmprint identification using featurelevel fusion. Pattern Recognit. 39(3), 478–487 (2006).CrossRefzbMATHGoogle Scholar
 28.J Wei, W Jia, H Wang, DF Zhu, in Emerging Intelligent Computing Technology and Applications. Improved competitive code for palmprint recognition using simplified gabor filter, (2009), pp. 371–7. ISBN:3642040691 9783642040696.Google Scholar
 29.Z Khan, A Mian, Y Hu, in IEEE International Conference on Computer Vision. Contour code: Robust and efficient multispectral palmprint encoding for human recognition (Barcelona, 2011), pp. 1935–1942.Google Scholar
 30.A Kumar, D Zhang, Personal authentication using multiple palmprint representation. Pattern Recognit. 38(10), 1695–1704 (2005).CrossRefGoogle Scholar
 31.W Li, J You, D Zhang, Texturebased palmprint retrieval using a layered search scheme for personal identification. Multimedia, IEEE Trans. 7(5), 891–8 (2005).CrossRefGoogle Scholar
 32.A Kong, D Zhang, M Kamel, A survey of palmprint recognition. Pattern Recognit. 42(7), 1408–18 (2009).CrossRefGoogle Scholar
 33.D Zhang, W Zuo, F Yue, A comparative study of palmprint recognition algorithms. ACM Comput. Surv. 44(1), 2–1237 (2012).CrossRefGoogle Scholar
 34.J Kannala, E Rahtu, in Pattern Recognition (ICPR), 2012 21st International Conference On. Bsif: Binarized statistical image features (IEEE,Tsukuba, 2012), pp. 1363–6.Google Scholar
 35.R Raghavendra, C Busch, in Proceedings of the 2Nd ACM Workshop on Information Hiding and Multimedia Security, IH&MMSec ’14. Robust palmprint verification using sparse representation of binarized statistical features: a comprehensive study (ACM,New York, NY, USA, 2014), pp. 181–5.Google Scholar
 36.D Zhang, WK Kong, J You, M Wong, Online palmprint identification. Pattern Anal. Mach. Intell. IEEE Trans. 25(9), 1041–50 (2003).CrossRefGoogle Scholar
 37.A Kumar, in Sixth Indian Conference on Computer Vision, Graphics Image Processing (ICVGIP). Incorporating cohort information for reliable palmprint authentication (Bhubaneswar, 2008), pp. 583–90.Google Scholar
 38.R Raghavendra, C Busch, Novel image fusion scheme based on dependency measure for robust multispectral palmprint recognition. Pattern Recognit. 44(6), 2505–221 (2014).Google Scholar
 39.H Lee, C Ekanadham, AY Ng, in Advances in Neural Information Processing Systems. Sparse deep belief net model for visual area v2, (2008), pp. 873–80. www.eecs.umich.edu.Google Scholar
 40.G Hinton, S Osindero, YW Teh, A fast learning algorithm for deep belief nets. Neural Comput. 18(7), 1527–54 (2006).MathSciNetCrossRefzbMATHGoogle Scholar
 41.P Vincent, H Larochelle, Y Bengio, PA Manzagol, in Proceedings of the 25th International Conference on Machine Learning. Extracting and composing robust features with denoising autoencoders (ACM,New York, NY, USA, 2008), pp. 1096–1103.Google Scholar
 42.BA Olshausen, DJ Field, Sparse coding with an overcomplete basis set: A strategy employed by v1?Vision Res. 37(23), 3311–25 (1997).CrossRefGoogle Scholar
 43.JH van Hateren, A van der Schaaf, Independent component filters of natural images compared with simple cells in primary visual cortex. Proc. R. Soc. London. Series B: Biol. Sci. 265(1394), 359–66 (1998).CrossRefGoogle Scholar
 44.AJ Bell, TJ Sejnowski, The independent components of natural scenes are edge filters. Vis. Res. 37(23), 3327–38 (1997).CrossRefGoogle Scholar
 45.A Hyvarinen, J Hurri, PO Hoyer, Natural Image Statistics (Springer, Berlin, 2009).CrossRefGoogle Scholar
 46.J Mairal, F Bach, J Ponce, G Sapiro, Online learning for matrix factorization and sparse coding. J. Mach. Learn. Res. 11, 19–60 (2010).MathSciNetzbMATHGoogle Scholar
Copyright information
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.