Image Signature Verification The Tell-Tale Style of Form-Characteristics
  • Print
  • Share

Automatic Signature Verification for Wet Ink Signatures

Signatures scanned from Checks and other Documents – Processing of Signature Images

The core capability for image signature verification in SOFTPRO's products is based on the SIVAL engine, a neural-net-based engine originally developed and trained by IBM in their German neural-net laboratories in 1989 - 1994. This was achieved in partnership with SOFTPRO, a Swiss partner of SOFTPRO, and a Swiss bank.Today SIVAL is fully owned and maintained by SOFTPRO. The SIVAL engine is build into the SignPlus ASV engine and the SignWare library, where it is part of the biometric signature combo verification of both static and dynamic signature parameters.

SIVAL is looking for many different Form Characteristics, such as ... zoom SIVAL is looking for many different Form Characteristics, such as ...

zoom image: SIVAL is looking for many different Form Characteristics, such as ...

SIVAL is looking for many different Form Characteristics, such as ...

SIVAL has an eye for form characteristics that are also relevant for a forensic expert. If it can be extracted and calculated by computer algorithms, then it can be a factor in the neural net training. Of course only so far as the information is avialble in the image. The boundary conditions for verifications from images are equal both for human and machine. No stereoscopic, three-dimensional aspects, just black picsels versus white picsels. Make up your mind! Quick!

In order to achieve automatic verification results that create benefit for one’s operations, it is necessary to optimize the signature capture environments. I.e., to achieve optimal results from the employment of any signature verification engine, it is important to provide a pair of signatures to the verification engine that is free from image related challenges. In most real life environments “free from challenges” is not achievable. Nevertheless, eventually it remains crucial for the quality of verification results to properly address such issues to at least minimize that challenge. The SignPlus Product Suite comes with a number of pre-processing features that deliberately address issues as known from the context of cheque processing environments. These features have to be carefully adjusted to the production environments at hand.