By ,
February 19, 2013 - 3:37 PM
When it comes to forms, vote-by-mail applications, checks or other legal documents, automated signature verification technology can help to confirm the authenticity of signatures, and the legitimacy of information on documents being processed. Ideally, signature verification technology should offer a highly reliable forgery detection system, enable more documents to be processed, and call on human intervention for exception handling only. The results should surpass manual-only verification, while streamlining current processes, and adjust to the specific needs of an organization or business unit.
So what technical details should you consider when evaluating automated signature verification systems?
Allows for multiple reference signatures
Comparing signatures against multiple reference signatures reduces the number of false positives by taking into consideration random deviations inherent to human handwriting. The major problem facing signature verification technology is differentiating between the consistent and unstable parts of the signature that vary with each signing. Therefore, the signature verification process has to have a certain tolerance, but it also must be sensitive enough to pick out distinctive evidence of a forgery, the proverbial needle in the haystack.
Automated signature verification systems must consider more data to detect the stable distinctive characteristics in a signature, and to focus on them in the verification process simultaneously, ignoring random distortions and variations inherent in genuine signatures. The systems should allow the update of reference signatures and add new ones to ensure the availability of complete and current data for verification. And when more than one reference prototype is used for comparison, automated signature verification systems should indicate which of them was found to best match the signature in question.
Provides confidence values
Assigning confidence values gives users the ability to plan different actions based on custom interpretations of the results. The confidence value indicates how certain the verification process is about the match between the signature presented and the corresponding reference signature. Depending on the confidence value chosen as a threshold, it is possible to change the percentage of “false accepts” and “false rejects” for a specific application.
Support diverse/multiple output options
Applications that may benefit from signature verification are diverse and so are the requirements, purposes, and scenarios that have to be implemented for each. Automated signature verification should account for multiple types of output results. It should also offer the flexibility and adaptability to be integrated in any environment.
Preprocessed images prior to recognition
Image preprocessing is an important stage that, to a large extent, predetermines the results of verification. Both insufficient cleaning, as well as the removal of significant pieces of a signature can cause unreliable results and increase the rate of false alarms. The task of image preprocessing is, in many cases, no less challenging than the verification itself. Nonetheless, removing lines, preprinted text, intrusions from other fields, stamps and other undesired elements is an important first step to leave a clean signature image for verification.
Automatic location of the signature
Automated location of signatures on checks or documents is an important, and often complicated task. State-of-the-art signature verification systems can automatically locate one or two signatures on checks and IRD images. They also can be configured to locate signatures on documents and forms.
Tatiana Vazioulina is Sr. Product Marketing Manager at Parascript, a company that develops cursive, handprint and machine print recognition technology, enabling business automation in areas like forms processing, check processing and fraud prevention, online at www.parascript.com. For more information, please visit: http://info.parascript.com/automating-signature-verification-video
Rate Post
You need to log in to rate blog posts.
Click here to login.
This post and comment(s) reflect the personal perspectives of community members, and not necessarily those of their employers or of AIIM International
