Gautam Kumar
Sambit Bakshi
Skip Abstract Section
Abstract
The growing integration of technology into our lives has resulted in unprecedented amounts of data that are being exchanged among devices in an Internet of Things (IoT) environment. Authentication, identification, and device heterogeneities are major security and privacy concerns in IoT. One of the most effective solutions to avoid unauthorized access to sensitive information is biometrics. Deep learning-based biometric systems have been proven to outperform traditional image processing and machine learning techniques. However, the image quality covariates associated with blur, resolution, illumination, and noise predominantly affect recognition performance. Therefore, assessing the robustness of the developed solution is another important concern that still needs to be investigated. This article proposes a periocular region-based biometric system and explores the effect of image quality covariates (artifacts) on the performance of periocular recognition. To simulate the real-time scenarios and understand the consequences of blur, resolution, and bit-depth of images on the recognition accuracy of periocular biometrics, we modeled out-of-focus blur, camera shake blur, low-resolution, and low bit-depth image acquisition using Gaussian function, linear motion, interpolation, and bit plan slicing, respectively. All the images of the UBIRIS.v1 database are degraded by varying strength of image quality covariates to obtain degraded versions of the database. Afterward, deep models are trained with each degraded version of the database. The performance of the model is evaluated by measuring statistical parameters calculated from a confusion matrix. Experimental results show that among all types of covariates, camera shake blur has less effect on the recognition performance, while out-of-focus blur significantly impacts it. Irrespective of image quality, the convolutional neural network produces excellent results, which proves the robustness of the developed model.
- [1] . 2018. Experiments with ocular biometric datasets: A practitioner’s guideline. IT Profess. 20, 3 (2018), 50–63.
DOI : Google Scholar
Digital Library
- [2] . 2017. Internet of things security: A survey. J. Netw. Comput. Applic. 88 (2017), 10–28.
DOI: Google ScholarDigital Library - [3] . 2007. A comparative study of fingerprint image-quality estimation methods. IEEE Trans. Inf. Forens. Secur. 2, 4 (2007), 734–743.
DOI: Google ScholarDigital Library - [4] . 2017. The challenge of test data quality in data processing. J. Data Inf. Qual. 8, 2 (2017), 1–4.
DOI: Google ScholarDigital Library - [5] . 2021. Deep learning for pedestrian collective behavior analysis in smart cities: A model of group trajectory outlier detection. Inf. Fusion 65 (2021), 13–20.
DOI: Google ScholarCross Ref
- [6] . 2014. Iterative methods for image restoration. In Academic Press Library in Signal Processing, Vol. 4. Elsevier, 193–247. DOI: Google ScholarCross Ref
- [7] . 2000. The OpenCV library. Dr. Dobb’s J. Softw. Tools (2000). https://docs.opencv.org/3.4/da/d54/group__imgproc__transform.html#gga5bb5a1fea74ea38e1a5445ca803ff121acf959dca2480cc694ca016b81b442ceb.Google Scholar
- [8] . 2017. Deep learning-based crack damage detection using convolutional neural networks. Comput.-Aid. Civil Infrast. Eng. 32, 5 (2017), 361–378.
DOI: Google ScholarDigital Library - [9] . 1986. Stochastic sampling in computer graphics. ACM Trans. Graph. 5, 1 (1986), 51–72.
DOI :Google ScholarDigital Library - [10] . 2019. Authentication and authorization for mobile IoT devices using biofeatures: Recent advances and future trends. Secur. Commun. Netw. (2019).
DOI: Google ScholarDigital Library - [11] . 2013. A survey of Gaussian convolution algorithms. Image Process. (2013), 286–310.
DOI :Google ScholarCross Ref - [12] . 2002. Digital image processing. Prentice-Hall, Inc. (2002). Google ScholarDigital Library
- [13] . 2017. Strengths and weaknesses of deep learning models for face recognition against image degradations. IET Biomet. 7, 1 (2017), 81–89.
DOI :Google ScholarCross Ref - [14] . 2007. Performance of biometric quality measures. IEEE Trans. Pattern Anal. Mach. Intell. 29, 4 (2007), 531–543.
DOI: Google ScholarDigital Library - [15] . 2019. Study on the influence of variable stride scale change on image recognition in CNN. Multim. Tools Applic. 78, 21 (2019), 30027–30037.
DOI: Google ScholarCross Ref - [16] . 2009. Analysis of the utility of classical and novel speech quality measures for speaker verification. In 3rd International Conference on Advances in Biometrics (ICB’09). Springer-Verlag, Berlin, 434–442.
DOI: Google ScholarDigital Library - [17] . 2016. Deep residual learning for image recognition. In IEEE Conference on Computer Vision and Pattern Recognition. 770–778.
DOI :Google ScholarCross Ref - [18] . 2016. Low-resolution convolutional neural networks for video face recognition. In 13th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS). 221–227.
DOI :Google ScholarCross Ref - [19] . 2021. Fuzzy-in-the-loop-driven low-cost and secure biometric user access to server. IEEE Trans. Reliab. 70, 3 (2021), 1014–1025.
DOI: Google ScholarCross Ref - [20] . 2002. An evaluation of iris segmentation algorithms in challenging periocular images. Handbook of Iris Recognition. Springer.
DOI: Google ScholarCross Ref - [21] . 2016. How image degradations affect deep CNN-based face recognition? In International Conference of the Biometrics Special Interest Group (BIOSIG). 1–5.
DOI :Google ScholarCross Ref - [22] . 2012. ImageNet classification with deep convolutional neural networks. In 25th International Conference on Neural Information Processing Systems (NIPS). Curran Associates Inc., 1097–1105.
DOI :Google ScholarCross Ref - [23] . 2013. Applied Predictive Modeling, Vol. 26. Springer.Google ScholarCross Ref
- [24] . 2020. Non-overlapped blockwise interpolated local binary pattern as periocular feature. Multim. Tools Applic. (2020), 1–33.
DOI :Google ScholarDigital Library - [25] . 2009. Basic methods for image restoration and identification. In The Essential Guide to Image Processing. Elsevier, 323–348.
DOI: Google ScholarCross Ref - [26] . 2017. Embedded real-time fall detection using deep learning for elderly care. arXiv preprint arXiv:1711.11200 (2017).Google Scholar
- [27] . 2017. Noisy ocular recognition based on three convolutional neural networks. Sensors 17, 12 (2017), 2933.
DOI: Google ScholarCross Ref - [28] . 2019. Deep residual CNN-based ocular recognition based on rough pupil detection in the images by NIR camera sensor. Sensors 19, 4 (2019), 842.
DOI: Google ScholarCross Ref - [29] . 2021. Deep learning in security of internet of things. IEEE Internet Things J. (2021), 1–1.
DOI: Google ScholarCross Ref - [30] . 2021. Security and privacy techniques in IoT environment. Sensors 21, 1 (2021).
DOI: Google ScholarCross Ref - [31] . 2021. Medical image classification based on an adaptive size deep learning model. ACM Trans. Multim. Comput. Commun. Applic. 17, 3s (
Oct. 2021).DOI: Google ScholarDigital Library - [32] . 2020. Security of internet of things edge devices. Softw.: Pract. Exper. 51 (
02 2020).DOI: Google ScholarCross Ref - [33] . 2015. Internet of things (IoT): A literature review. J. Comput. Commun. 3, 05 (2015), 164–173.
DOI: Google ScholarCross Ref - [34] . 2009. Overview and framework for data and information quality research. J. Data Inf. Qual. 1, 1 (
June 2009).DOI: Google ScholarDigital Library - [35] . 2019. A privacy-enhancing framework for internet of things services. In International Conference on Network and System Security. Springer, 77–97.
DOI: Google ScholarDigital Library - [36] . 2017. On the effects of low-quality training data on information extraction from clinical reports. J. Data Inf. Qual. 9, 1 (2017), 1–25.
DOI: Google ScholarDigital Library - [37] . 2010. Performance evaluation of local appearance based periocular recognition. In 4th IEEE International Conference on Biometrics: Theory, Applications and Systems (BTAS).
DOI : Google ScholarCross Ref - [38] . 2016. Robust periocular recognition by fusing sparse representations of color and geometry information. J. Sig. Process. Syst. 82, 3 (
01 Mar. 2016), 403–417.DOI: Google ScholarDigital Library - [39] . 2021. A survey of iris datasets. Image Vis. Comput. 108 (2021), 104109.
DOI : Google ScholarCross Ref - [40] . 2012. Periocular recognition: Analysis of performance degradation factors. In 5th IAPR International Conference on Biometrics (ICB). 439–445.
DOI : Google ScholarCross Ref - [41] . 2013. Compensating for pose and illumination in unconstrained periocular biometrics. Int. J. Biomet. 5, 3-4 (2013), 336–359.
DOI : Google ScholarCross Ref - [42] . 2018. Correlation analysis of performance measures for multi-label classification. Inf. Process. Manag. 54, 3 (2018), 359–369.
DOI : Google ScholarCross Ref - [43] . 1983. Modeling motion blur in computer-generated images. ACM SIGGRAPH Comput. Graph. 17, 3 (1983), 389–399.
DOI : Google ScholarDigital Library - [44] . 2020. Evaluation: From precision, recall and F-measure to ROC, informedness, markedness and correlation. arXiv preprint arXiv:2010.16061 (2020).Google Scholar
- [45] . 2005. UBIRIS: A noisy iris image database. In Image Analysis and Processing – ICIAP. Springer, 970–977.
DOI : Google ScholarDigital Library - [46] . 2016. Visible-wavelength iris/periocular imaging and recognition surveillance environments. Image Vis. Comput. 55 (2016), 22–25.
DOI: Google ScholarCross Ref - [47] . 2010. The UBIRIS.v2: A database of visible wavelength iris images captured on-the-move and at-a-distance. IEEE Trans. Pattern Anal. Mach. Intell. 32, 8 (2010), 1529–1535.
DOI: Google ScholarDigital Library - [48] . 2020. Collaborative representation of blur invariant deep sparse features for periocular recognition from smartphones. Image Vis. Comput. 101 (2020), 103979.
DOI: Google ScholarCross Ref - [49] . 2016. Competition on mobile ocular biometric recognition. In IEEE International Conference on Image Processing (ICIP). 320–324.
DOI: Google ScholarCross Ref - [50] . 2018. Multi-frame super resolution for ocular biometrics. In IEEE Conference on Computer Vision and Pattern Recognition Workshops. 453–461.
DOI: Google ScholarCross Ref - [51] . 2017. Deep Learning for Computer Vision with Python: ImageNet Bundle. PyImageSearch. Google Scholar
- [52] . 2018. Effects of degradations on deep neural network architectures. arXiv preprint arXiv:1807.10108 (2018).Google Scholar
- [53] . 2017. Dynamic routing between capsules. In International Conference on Advances in Neural Information Processing Systems. 3856–3866.Google Scholar
- [54] . 2014. Fusing iris and periocular information for cross-sensor recognition. Pattern Recog. Lett. (2014).
DOI: Google ScholarDigital Library - [55] . 2012. Minutiae from bit-plane sliced thermal images for human face recognition. In International Conference on Soft Computing for Problem Solving (SocProS’11). Springer, 113–124.
DOI: Google ScholarCross Ref - [56] . 2002. Machine learning in automated text categorization. ACM Comput. Surv. 34, 1 (2002), 1–47.
DOI: Google ScholarDigital Library - [57] . 2014. On cross spectral periocular recognition. In IEEE International Conference on Image Processing (ICIP). 5007–5011.
DOI: Google ScholarCross Ref - [58] . 2014. Very deep convolutional networks for large-scale image recognition. arXiv Preprint arXiv:1409.1556 (2014).Google Scholar
- [59] . 2005. ROCR: Visualizing classifier performance in R. Bioinformatics 21, 20 (2005), 3940–3941.
DOI: Google ScholarDigital Library - [60] . 2013. Face recognition using PCA and bit-plane slicing. In 3rd International Conference on Trends in Information, Telecommunication and Computing. Springer, 515–523.
DOI: Google ScholarCross Ref - [61] . 2019. Data sharing and privacy for patient IoT devices using blockchain. In International Conference on Smart City and Informatization. Springer, 334–348.
DOI: Google ScholarCross Ref - [62] . 2015. Going deeper with convolutions. In IEEE Conference on Computer Vision and Pattern Recognition. 1–9.
DOI: Google ScholarCross Ref - [63] . 2018. Survey of authentication and authorization for the internet of things. Secur. Commun. Netw. (2018).
DOI: Google ScholarDigital Library - [64] . 2020. Person identification using fusion of iris and periocular deep features. Neural Netw. 122 (2020), 407–419.
DOI: Google ScholarDigital Library - [65] . 2011. Patch-based probabilistic image quality assessment for face selection and improved video-based face recognition. In IEEE Biometrics Workshop, Computer Vision and Pattern Recognition (CVPR) Workshops. 81–88.
DOI: Google ScholarCross Ref - [66] . 2010. On the fusion of periocular and iris biometrics in non-ideal imagery. 20th International Conference on Pattern Recognition (ICPR). 201–204.
DOI: Google ScholarDigital Library - [67] . 2020. A stride-based convolution decomposition method to stretch CNN acceleration algorithms for efficient and flexible hardware implementation. IEEE Trans. Circ. Syst. I: Reg. Pap. 67, 9 (2020), 3007–3020.
DOI: Google ScholarCross Ref
Index Terms
- Experimental Evaluation of Covariates Effects on Periocular Biometrics: A Robust Security Assessment Framework
Recommendations
Periocular recognition: how much facial expressions affect performance?
Using information near the human eye to perform biometric recognition has been gaining popularity. Previous works in this area, designated periocular recognition, show remarkably low error rates and particularly high robustness when data are acquired ...
Periocular Biometrics in the Visible Spectrum
The term periocular refers to the facial region in the immediate vicinity of the eye. Acquisition of the periocular biometric is expected to require less subject cooperation while permitting a larger depth of field compared to traditional ocular ...
Fusion of iris and periocular biometrics for cross-sensor identification
CCBR'12: Proceedings of the 7th Chinese conference on Biometric RecognitionAs a reliable personal identification method, iris recognition has been widely used for a large number of applications. Since a variety of iris devices produced by different vendors may be used for some large-scale applications, it is necessary to match ...
Comments