Skip to content

AI Innovation

Both the financial and healthcare industries are undergoing an AI evolution. Review our vision for Artificial Neural Networks (ANN) and deep learning targeting these industries.

AI and Deep Learning-Based Check Processing

Check recognition & fraud detection are the most important components in today's check processing and omni-channel capture. Learn how OrboAnywhere using OrbNet AI technology reduces costs and mitigates risk for any check image capture workflow.

AI-Based Healthcare Electronification

OrboAccess, powered by OrbNet AI, provides electronification to remits and payments, enabling RCM companies to automate posting, improve research, and deliver business intelligence

About Us

Celebrating 25 years of innovation, OrboGraph has transformed into an AI company delivering targeted automation solutions to the banking and healthcare industries.

Resources

From news and events to case studies, trends, and videos, this section provides a range of information resources for payment automation in the banking and healthcare industries.

Blogs

OrboGraph produces four blog series on a weekly basis covering topics from check processing, fraud prevention, AI technologies, RCM, and healthcare electronification. Select one the blog to the right. We hope you enjoy!

  • What is Forensic Document Examination?
  • Comparing Forensic AI technology to the Forensic Document Examiner role
  • Benefits of forensic document examination for fraud detection

Forensic document examination is foundational to OrboGraph’s OrbNet Forensic AI technology. To assist in describing the applicability of forensics in banking, fraud detection, and payments, we talked to forensic expert and author, Khody Detwiler, for this week’s Modernizing Omnichannel Check Fraud Detection post.

Khody Detweiler

Khody Detwiler

Forensic Document Examiner, Lesnevich & Detwiler

Focus on Forensic Document Examination

Joe Gregory, VP of Marketing at OrboGraph, interviewed Khody Detwiler for a discussion of forensic document examination and its impact on banking, fraud, and payments. Here are excerpts. (Click the button to the right to download entire interview.)

Joe Gregory:

Khody, thank you again for joining us. What really is forensic document examination?

Khody Detwiler:

Forensics as a whole is the application of science to law. Forensic document examination is a subcategory of the forensic field, dealing with the authenticity of the documents. In a way, it is similar to the ballistics or fingerprint forensic disciplines.

It could involve the examination of signatures, handwriting, a security feature – as in the situation with passports or travel documents – or identifying the age of a document. What we do as forensic document examiners is analyze documents to make determinations as to authenticity, which could be the document as a whole or some part of the document...

Joe Gregory:

What education is needed in order to practice as a forensic document examiner?

Khody Detwiler:

I went through a formal two-year training program that any qualified forensic document examiner will go through. This training is currently published by the Scientific Working Group for Forensic Document Examination (SWGDOC). You can visit their website to access all of the published standards. (https://www.swgdoc.org/)

A large part of the formal training program involves sitting alongside a senior examiner to review every case coming into the lab, complete reading assignments, initiate research, complete a cycle of working cases, and undergo a variety of tests and oral moot courts.

Joe Gregory:

What is the most interesting type of case that you were involved with so far?

Khody Detwiler:

The most interesting type of case — personally, I enjoy larger, complex cases that take a lot of time because law enforcement is investigating many moving pieces.

I was involved with was a case out of Saudi Arabia. There were over 100 international banks involved, and what we essentially found is that a group of identical signatures were being used over and over again, on different loan documents across these different banks. Total at stake was a little over $20 billion in that particular case...

Joe Gregory:

How can you really identify the differences between handwriting samples?

Khody Detwiler:

If you have two signatures that are pictorially different, before we can get to an opinion that one is not a genuine signature, you have to understand what those differences mean. So, to actually work a case and evaluate differences, we use a sampling of what we call known signatures, or standard signatures. It is similar to how Anywhere Fraud with OrbNet Forensic AI technology works, essentially comparing known standards to the signatures in question. 

As a forensic document examiner, I am evaluating whether or not these apparent differences are actually just part of a range of variation exhibited in the genuine signatures of the true author, or if they're actually differences that would be attributed to another writer.

When a difference is noted you have to evaluate — is that difference significant?...

Joe Gregory:

Can you talk briefly about what the 21 discriminating elements mean to signature analysis?

Khody Detwiler:

We have the two different groups that we as forensic document examiners have to look at: pictorial and execution. There are 21 discriminating elements that have been brought out in literature over the years. For each of those categories, you can have a variety of subcategories.

Part of the discriminating elements are considered to be components of style, or the pictorial features of the writing. For example, features like spacing, sizing, letter formations, are in this category...

Joe Gregory:

Can you compare and contrast AI software verses how a document examiner works?

Khody Detwiler:

There is a lot of commonality between how software operates and how a document examiner looks at documents and signatures. Tying back to those 21 discriminating elements, the software and examiner analyze similar features such as spacing, sizing, letter formation, etc.

AI software doesn’t exactly tell you which features are being analyzed by name. In the case of OrbNet Forensic AI, I know there are 512 feature vectors used in feature extraction which are incorporated with this image analysis. Also remember that software typically analyzes 200 dpi bitonal images in check processing.

Looking at the check as a whole, the software can certainly run the analysis faster and with greater depth and breadth...

Joe Gregory:

Do you think that the financial industry could use specific programs to help them for fraud review processes or court cases? How about interbank fraud disputes?

Khody Detwiler:

Absolutely. As a comparison, you can think of this process similar to examining travel documents at an airport. First, the TSA agent making the initial contact with the traveler is confronted with a phony travel document — that's essentially “tier one” or “screening step #1.”

They look at it, they'll flag it, that goes back to somebody else who has a little bit more training, a little bit more knowledge, and maybe some more sophisticated equipment than a UV light and a loupe. They then determine that  there's definitely a problem, but don't know exactly what it is or how it was done.

The document then gets escalated to “tier three,” which would be the forensic document laboratory...

We want to thank Khody for his time. It certainly was interesting to learn about forensic document examination and methodology. To get in contact with Khody Detwiler, please click here to connect with him on LinkedIn.

Lesnevich and Detwiler

Lesnevich & Detwiler

7550 Woodbury Pike
Roaring Spring, PA 16673

info@lesnevich.com

Leave a Comment





Review your needs with an OrboGraph expert.

Sign up below for your complimentary assessment or to request
estimated solution pricing from OrboGraph.