Effect of identity mapping, transfer learning and domain knowledge on the robustness and generalization ability of a network: a biometric based case study

  • Avantika SinghEmail author
  • Aditya Nigam
Original Research



This work evaluates two widely validated CNN models VGG and ResNet in the biometric domain based applications. These models have shown exceptional performance on the large ImageNet dataset but the predictive capability of these models for domain specific-tasks where limited data samples are present need to be checked. Here in this paper efforts have been made to analyze two interesting biometric problems (1) multi-class oculus classification (2) fingerprint sensor classification. The distinguish experimental results are evaluated on the benchmark datasets like IIIT-D, ND, IITK contact-lens datasets and FVC 2002, FVC 2004, FVC 2006, IIITD-MOLF and IITK fingerprint datasets. We have raised many interesting questions regarding the reliability and applicability of deep-learning models specifically for biometric based applications. Experimental results along with in-depth feature analysis have shown that indeed residual connections with pre-trained network provides good prior for model weights and thus helps in better generalization.

Graphical abstract


Deep learning Transfer learning Contact lens detection Fingerprint sensor classification Robustness analysis 



  1. Agarwal A, Singh R, Vatsa M (2016) Fingerprint sensor classification via mélange of handcrafted features. In: 23rd International conference on pattern recognition, ICPR, pp 3001–3006Google Scholar
  2. Ahuja K, Islam R, Barbhuiya FA, Dey K (2017) Convolutional neural networks for ocular smartphone-based biometrics. Pattern Recognit Lett 91:17–26CrossRefGoogle Scholar
  3. Bartlow N, Kalka ND, Cukic B, Ross A (2009) Identifying sensors from fingerprint images. In: IEEE conference on computer vision and pattern recognition, CVPR workshops, pp 78–84Google Scholar
  4. Cappelli R, Ferrara M, Franco A, Maltoni D (2007) Fingerprint verification competition 2006. Biom Technol Today 15:7–9CrossRefGoogle Scholar
  5. Daugman J (2002) How iris recognition works. In: Proceedings of the 2002 international conference on image processing, ICIP, pp 33–36Google Scholar
  6. Doyle J, Bowyer K, Flynn P (2013) Variation in accuracy of textured contact lens detection based on sensor and lens pattern. 6th IEEE international conference on biometrics: theory. applications and systems, BTAS, pp 1–7Google Scholar
  7. Gautam G, Mukhopadhyay S (2018) Contact lens detection using transfer learning with deep representations. In: 2018 International joint conference on neural networks, IJCNN, pp 1–8Google Scholar
  8. Grün F, Rupprecht C, Navab N, Tombari F (2016) A taxonomy and library for visualizing learned features in convolutional neural networks. CoRR arXiv:1606.07757
  9. He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In: Conference on computer vision and pattern recognition, CVPR, pp 770–778Google Scholar
  10. Kohli N, Yadav D, Vatsa M, Singh R (2013) Revisiting iris recognition with color cosmetic contact lenses. In: International conference on biometrics, ICB, pp 1–7Google Scholar
  11. LeCun Y, Haffner P, Bottou L, Bengio Y (1999) Object recognition with gradient-based learning. Shape Contour Group Comput Vis 1681:319–347CrossRefGoogle Scholar
  12. Lee EC, Park KR, Kim J (2006) Fake iris detection by using Purkinje image. In: International conference on biometrics, IAPR, pp 397–403Google Scholar
  13. Lovish Nigam A, Kumar B, Gupta P (2015) Robust contact lens detection using local phase quantization and binary gabor pattern. 16th International conference. Computer analysis of images and patterns, CAIP, pp 702–714Google Scholar
  14. Lukas J, Fridrich J, Goljan M (2006) Digital camera identification from sensor pattern noise. IEEE Trans Inf Forensics Secur 1(2):205–214CrossRefGoogle Scholar
  15. Maio D, Maltoni D, Cappelli R, Wayman JL, Jain AK (2002) Fvc2002: second fingerprint verification competition. In: 16th International conference on pattern recognition, vol 3, pp 811–814Google Scholar
  16. Maio D, Maltoni D, Cappelli R, Wayman J, Jain A (2005) Combining multiple matchers for fingerprint verification: a case study in FVC2004. 13th International conference. Image analysis and processing, ICIAP, pp 1035–1042Google Scholar
  17. Mhaskar H, Liao Q, Poggio TA (2017) When and why are deep networks better than shallow ones? In: Proceedings of the thirty-first AAAI conference on artificial intelligence, pp 2343–2349Google Scholar
  18. Raghavendra R, Raja KB, Busch C (2014) Ensemble of statistically independent filters for robust contact lens detection in iris images. In: Proceedings of Indian conference on computer vision graphics and image processing, ICVGIP, pp 241–247Google Scholar
  19. Raghavendra R, Raja KB, Busch C (2017) Contlensnet: Robust iris contact lens detection using deep convolutional neural networks. In: IEEE winter conference on applications of computer vision, WACV, pp 1160–1167Google Scholar
  20. Rattani A, Derakhshani R (2017) On fine-tuning convolutional neural networks for smartphone based ocular recognition. In: IEEE international joint conference on biometrics, IJCB, pp 762–767Google Scholar
  21. Rattani A, Reddy N, Derakhshani R (2018) Multi-biometric convolutional neural networks for mobile user authentication. In: IEEE international symposium on technologies for homeland security, HST, pp 1–6Google Scholar
  22. Ren S, He K, Girshick R, Sun J (2017) Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Trans Pattern Anal Mach Intell 39(6):1137–1149CrossRefGoogle Scholar
  23. Ring S, Bowyer K (2008) Detection of iris texture distortions by analyzing iris code matching results. In: IEEE international conference on biometrics: theory, applications, and systems, BTAS, pp 1–6Google Scholar
  24. Russakovsky O, Deng J, Su H, Krause J, Satheesh S, Ma S, Huang Z, Karpathy A, Khosla A, Bernstein MS, Berg AC, Li F (2015) Imagenet large scale visual recognition challenge. Int J Comput Vis 115(3):211–252MathSciNetCrossRefGoogle Scholar
  25. Sankaran A, Vatsa M, Singh R (2015) Multisensor optical and latent fingerprint database. In: IEEE access, pp 653–665Google Scholar
  26. Silva P, Luz E, Baeta R, Pedrini H, Falcao AX, Menotti D (2015) An approach to iris contact lens detection based on deep image representations. In: Graphics, patterns and images (SIBGRAPI), pp 157–164Google Scholar
  27. Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. CoRR arXiv:1409.1556
  28. Vlachos M, Dermatas E (2010) Multi-scale retinal vessel segmentation using line tracking. Comput Med Imaging Graph 34(3):213–227CrossRefGoogle Scholar
  29. Wolpert D, Macready W (1997) No free lunch theorems for optimization. IEEE Trans Evol Comput 1(1):67–82CrossRefGoogle Scholar
  30. Yadav D, Kohli N, Doyle JS Jr, Singh R, Vatsa M, Bowyer KW (2014) Unraveling the effect of textured contact lenses on iris recognition. IEEE Trans Inf Forensics Secur 9:851–862CrossRefGoogle Scholar
  31. Zeiler MD, Fergus R (2014) Visualizing and understanding convolutional networks. In: European conference on computer vision, ECCV, pp 818–833Google Scholar
  32. Zhang Zhang L, Zhang D, Zhu H (2011) Ensemble of local and global information for finger–knuckle-print recognition. Pattern Recognit 44(9):1990–1998CrossRefGoogle Scholar

Copyright information

© Springer-Verlag GmbH Germany, part of Springer Nature 2019

Authors and Affiliations

  1. 1.Indian Institute of Technology MandiMandiIndia

Personalised recommendations