Palmprint Biometric on Smartphones

Background

 In a world where smartphones are becoming ubiquitous and more powerful than ever, securing these devices has become a key issue. We are so invested in our smartphones that once they are compromised our personal lives are also laid bare to a cyber-criminal. Digital forgery and identity theft can be achieved in the blink of an eye, for this reason it becomes increasingly important that we can authenticate ourselves reliably.

Equipped with cameras of increasing capability, smartphones are prime candidates for image processing software designed to recognize objects and people around them. By recognizing biometric patterns such as iris, retina, hand vein or palmprint they can provide reliable user authentication.

Palmprints have been around for many years, and even if prehistoric humans only used them for decorative purposes in caves, as portrayed in Fig.1, they have always provided a mark of identity. In modern times palmprints are mostly encountered in their latent form, in the field of forensics, where they are used for identification purposes.

palmprints_cave

Figure 1. Hand paintings discovered in Spain, dated around 40,000 years ago. (source: https://anthropology.net/2012/06/14/were-paleolithic-european-cave-paintings-made-by-neanderthals/ accessed on 3.06.2016)

The information contained in the hand (see Fig.2), in the palm region is a good solution to control access problems and has been used for years now, with the help of hand scanning equipment – such as a Schlage scanner displayed in Fig.3 (a).

pp_structure
Figure 2. Structure of the palmprint region

The increasing computational power of handheld devices in general and smartphones in particular, coupled with high quality imaging systems makes these devices excellent candidates for implementation of biometrics based security solutions.

fig_3

Figure 3. a) Hand Scanner device (source http://store.amgtime.com/hardware/AMG-HandPunch-3000E accessed on 3.06.2016) b) Use case for capturing palmprint using a smartphone

By using the rear camera of the smartphone, a user can safely capture a good quality image of a hand, as displayed in Fig.3 (b), which can then be processed and used to authenticate the user.

Research Goals

The goal of this project is to develop an algorithm for automatic palmprint based authentication. The algorithm should be light enough to be easily implementable on a smartphone and robust when used in unconstrained conditions.

The challenges associated with an unconstrained acquisition of a hand include:

  • Orientation and scale normalization
  • Hand pose
  • Lighting conditions
  • Inter-device variation of the image quality

By using the landmarks shown in Fig.4 one is able to determine the rotation and reduce the influence of the variable scale. Lighting conditions can be overcome by specific feature extraction techniques and illumination invariant algorithms.

pp_landmarks

Figure 4. The circles’ centers are used to determine a surface from within the hand that can be extracted regardless of the hand’s rotation.

Check out the Palmprint Database(s) we have listed on this website – here!

Publications

  1. H. Javidnia, A. Ungureanu,  C. Costache,  P. Corcoran, “Palmprint as a smartphone biometric”(Inproceeding) 2016 IEEE International Conference on Consumer Electronics (ICCE), pp. 463–466, IEEE 2016.
  2. A. Ungureanu, H. Javidnia, C. Costache,  P. Corcoran, “A review and comparative study of skin segmentation techniques for handheld imaging devices” (Inproceeding) 2016 IEEE International Conference on Consumer Electronics (ICCE), pp. 530–531, IEEE 2016.
  3. H. Javidnia, A. Ungureanu, P. Corcoran, “Palm-print recognition for authentication on smartphones” (Inproceeding) 2015 IEEE International Symposium on Technology and Society (ISTAS), pp. 1–5, IEEE 2015.
  4. A.-S. Ungureanu, S. Thavalengal, T. E. Cognard, C. Costache, and P. Corcoran, “Unconstrained palmprint as a smartphone biometric,” IEEE Trans. Consum. Electron., vol. 63, no. 3, pp. 334–342, Aug. 2017.
  5. A. Ungureanu, S. Bazrafkan, and P. Corcoran, “Deep Learning for hand segmentation in complex backgrounds,” pp. 604–605, 2018.