The Research On Embedded Liveness Detection Palmprint Recognition System

As a kind of biometric technology, palmprint recognition has the following advantages: containing more information, strong anti-jamming and can extract feature with low-resolution image. Therefore, it has a broad application prospects. However, PC-based palmprint has following disadvantages: large size, high cost and inconvenient to carry.Besides, with biometric identification systems are widely used, a variety of methods that can be used to attack or fool biometric recognition systems. So liveness detection must be used to increase the security and reliability of systems, which can effectively prevent or avoid fraud and attack recognition system.Through analysis of the advantages and disadvantages of a variety of embedded systems and based on the latest extensive discourse and technology in liveness detection field, we design an embedded liveness detection palmprint recognition system. Main contents are as follows:The building of the hardware platform for embedded palmprint system. The system includes peripherals and embedded engines. Embedded engine is based on DSP + FPGA, including acquisition module, storage module, input and output module. To fully utilize the advantages of the two kinds of processors, DSP is used to complete the core functions of palmprint recognition algorithms and FPGA is used to be a general-purpose processor to control the whole system.The development of drivers and control software for embedded palmprint system. Design and develop the drivers of keyboard interface, LCD display, SD card memory modules, FAT file system and control software of the system. Based on the analysis of existing palmprint recognition algorithm, combined with the characteristics of embedded platform, transplant, optimize or re-implementation the algorithm and develop software packages.Conducting research on palmprint liveness detection technology and applying it to the new system. Liveness detection can be divided into two stages that are signal sampling and processing. According to the palm characteristics and the our team's real situation, we combine the two kind of methods: in the signal acquisition stage, using penetration characteristics of near-infrared light to the palm skin, as well as the absorption characteristics of palm vein to do liveness detection and in the signal processing stage, using pattern classification approach, by analyzing the image texture and brightness to do liveness detection.Test results show that the embedded system not only achieve function of the palmprint system, but also reduces the cost and size, lower the power consumption, improve efficiency, enhance the stability and reliability and the liveness detection elements greatly enhance the anti-counterfeiting capabilities.