KTM for Invoice Processing.
The Kofax Transformation Modules is an excellent addition to the Kofax Capture family when processing invoices. The product is able to “Read” the major invoice values from the invoice image accurately and with confidence. It has the ability to “Learn” where on the invoice image the values are located so its accuracy can be increased over time.
The major invoice values that KTM can read and “Learn” out of the box are:
The accuracy out of the box is very low since the software needs to be “Trained” on sample invoices to learn where the values are located. It does this by layout classification so it can handle different vendor’s invoices and learn to accurately read the values.
In addition, there are lots of options for the builder of the Project to help the extraction of the values. These options include objects like Locators, Formatting, Key Word matching, and Validation. Through a combination of these options, just about any invoice can be read and processed with little user correction required.
However, it takes a lot of work to “tune” the Locators and Key Word matching to get the initial accuracy to as high a rate as possible. This is more an art then a science and therefore requires a person with intimate knowledge of KTM and all it’s abilities. Also, knowledge of basic script programming is required since there are things that cannot be done by KTM otherwise. ImageSource can provide such people that have the knowledge and experience to “tune” KTM to be as accurate as can be done with the current OCR technology.
Sr. Systems Engineer