Distributed infrared biometric sensing for lightweight human identification systems

This paper reports the findings of a new distributed infrared biometric sensing approach in lightweight human identification. Pyroelectric infrared (PIR) sensor arrays are employed in infrared biometric sensing while the vector quantization (VQ) based algorithm is developed for identifying walkers in the context of the path-constrained walking. The main contribution of this paper is twofold. First, a distributed sensing structure is used to organize two modules of PIR sensor arrays for capturing walkers´ biometric characteristics. The advantage of using the distributed sensing paradigm is that the synergistic motion patterns of upper- and -lower limbs can be encoded and thus the more discriminative features can be decoded accordingly. This is the new consideration of using PIR sensors in building a lightweight biometric sensing system. Second, VQ is introduced in modeling the features of the episodes of walkers´ limb movement with code books, which contributes to performance improvement in human identification. Experimental studies are conducted to validate the proposed method.