Biometric authentication is now widely utilised as an alterna-tive to passwords on IoT devices such as smartphones. However, present biometric systems are subject to spoofing attacks. Several liveness detec-tion techniques have been presented to determine whether a live person or an artificial duplicate is in front of the biometric sensor. However, the challenge remains unsolved due to the difficulty in identifying discrimina-tive and computationally affordable traits for spoofing attacks. Further-more, previous liveness detection techniques are not particularly oriented towards mobile biometrics, making them mostly unsuitable for portable devices. As a solution, we created a software-based multi-biometric pro-totype that detects face, iris, and fingerprint spoofing attacks on mo-bile devices. We present Mobile Biometric Liveness Detection techniques (MBLDT). Apart from the fact that conventional mobile devices perform badly for floating point applications, MBLDT is computed in linear time with respect to the amount of pixels and does not require floating point computation. As a result, our technique is solely simple, quick, and ef-ficient, making it ideal for mobile devices. Furthermore, unlike previous approaches, our method effectively detects liveness using the same lone image descriptor technique for three biometric features, namely face, iris, and fingerprint. Furthermore, our system detects liveness using only one image, which can also be used for recognition. Experiments with real spoofing attacks on widely available face, iris, and fingerprint data sets have yielded encouraging results.
