QR Code Based Signature Encoding System

Abstract

Signature on a paper is a most general authentication method used in the world. Personal signature plays significant role in a place like Bank in money transactions. A signature encoding system provides secure reliable and fast money transaction. Banks need to handle thousands of transactions in a second. Therefore signature should encode to very simple form. QR code would be one of feasible solution in that context. QR code could use as simple encoded symbol that represent signature. This paper discusses a QR code-based signature encoding system implemented with an FPGA based hardware platform to increase the speed of encoding. The image of the signature is used to generate QR code. Direct generation QR code from the webcam image is not possible. Therefore image is pre-compressed before encoding. First the input image is normalized and its contrast is enhanced by using histogram equalization. It is converted into a binary image and Binary image matrix is compressed into a simple matrix using two-dimensional wavelet analysis.Two matrixes are produce as output. One matrix consists of coefficient information and other matrix consists mapping detail. The matrix consist of mapping detail is converted to a QR code. Matrix which consist coefficients is stored as secondary information for decoding process. The system is implemented in FPGA based embedded system with a webcam interface. By using this method the signatures are encoded with accuracy around 60% and about 65% differentiation accuracy of two different signatures were observed