Innovative CountErspoofing for fingerprint bIOmetric seNsor – INCEPTION

Abstract

Biometric solutions increasingly appear into our everyday life, such as for instance access control terminals, biometric passports and mobile phones, with fingerprint recognition system in the first place. This continuous expansion of biometrics leads to attempts to circumvent biometric systems. "Spoofing attacks" or lures such as the use of false fingers for fingerprint sensors are a real threat and resistance to fakes is becoming a critical issue concerning the future of this technology. More particularly, thin films, made of transparent material like gelatine, latex of silicone paste are very difficult to detect by current sensors.

The goal of INCEPTION is to develop an innovative technique to counteract the use of spoofing attacks against fingerprint biometric sensors based on a recent technique, initially developed for medical purposes: Diffuse Reflectance Spectroscopy (DRS). This technique allows the measurement of two optical properties of tissues: light absorption coefficient and light diffusion coefficient. The principle of the DRS fingerprint sensor is to enlighten the finger with a narrow and multispectral beam of light, and measure the intensity of transmitted light at several distances from the source to estimate the optical properties of the finger at different depths. The different optical properties of the multiple layer of the tissue will be estimated by taking advantage of the high number of spatially distributed light detectors (sensitive pixels). By estimating a different absorption and diffusion value in the tissue layers (stratum corneum, epidermis and dermis) we will improve the detection of thin transparent films and open the way to severed limb detection. Two complementary strategies will be developed and tested: finding of the optimal illumination pattern and multilayer estimation of the optical properties of the finger.
Therefore, INCEPTION have the following objectives: (1) design and manufacture the prototype of a new fingerprint sensor relying on DRS on a Thin Field Transistor (TFT) matrix for liveness detection, (2) develop a new algorithm to gain access to a multilayer estimation of the optical properties of the finger, (3) develop spoof decision algorithm for this new sensor, (4) test and evaluate the prototype on a representative number of lures and real fingers and (5) provide a detailed state of the art of publications and IP of optical based liveness detection for fingerprint sensors and use it to set up an exploitation strategy.
Besides, during the research and development activities of the project, ethical and privacy issues will be considered with care regarding the management of sensitive personal data. Finally, industrial benefits are expected out of the INCEPTION project that are mainly related to the design and medium term availability of highly robust fingerprint sensors to lures. If successful, this project will impose a new standard of robustness for biometry sensors, consequently improving citizen’s security and trust, having a huge impact at the society level.