Handwritten Signature Confidentiality through Segmented Curve Generation
Ebenezer R. H. P. Isaac, J. Visumathi · 2014
Abstract — While Handwritten Signature Verification (HSV) techniques thwart attacks on the security associated with authentication, the signature itself may be prone to attacks with respect to data security. Hence it is vital to store the signatures as encrypted parameters. But storage of the signature as parameters of features would mean that it would not be possible to retrieve the original signature image for cross referencing in case of any error such as a false rejection. Though image encryption may provide end-to-end security, it requires the image to be reproduced at least temporarily on to the memory of the system on which the signature verification is to be performed. This leads to the vulnerability of what is called an image shadowing attack also known as a dead-forensic attack. This article poses a technique that allows the HSV system to store the original reference signatures as a set of encrypted curve equations providing added confidentiality to the signatures as well as facilitating the signatures to be retrievable in case of any discrepancies in the signature verification algorithm.