Forms Identification and Alignment Library

     buy on-line   download demo or evaluation version    Recogniform Reader     info     

This engine allows to perform the automatic identification and classification of forms, i.e. allows to recognize an classify a form among different forms previously defined in a set of possible forms. It releases the following data:

• Template ID;
• Template description;
• % of similarity with template.

For example, l et's suppose to have a collection of forms made of different classes (different kind of forms): we can process a mixed batch of documents simultaneously, and the recognition engine will release, for each of them, its "class".

In this way you can order and identify forms in a very fast and simple way.

This library also allows to perform the alignment of the form, according to its template. This means that the system automatically determines all differences between the form and the template, and that make form and template "different":

• horizontal and vertical offset;
• horizontal and vertical stretch;
• orientation;
• skew;
• resolution.

With these info it's possible to identify with great accuracy the form areas of interest on a specific form, even if they've measured on a sample form (template).

For example, an area in a XY position on a form, may not be in the same position on ALL forms scanned, even if they're all of the same kind: this because of stretch, offset or orientation defects.

Concluding, the operation of form alignment according to its template is a very important operation to perform, and mandatory to perform any identification anda data-capture operation.

This new release of Form Identification library has been totally re-developed, and does not use the Neural Network technology used in the previous release, so does not require a previous "learning" process. This leads to many advantages:

• don't need many samples of a form: just need one;
• in every moment you can add a template to the current set of forms.

To make an example, we could use these 3 tremplates  as SET:

Process this form...

...it's correctly identified as belonging to the class of the second teplate (9452), even if there's a consistent horizontal and vertical offset generated from the scanning activity.

 

If we process this form...

...the form is 180° oriented and correctly identified as belonging to the class of the third template (SS4), even if it has been captured upside-down.

In both cases offset and orientation info reported allow to identify areas correctly from a data-capture application that would process it.

Main Features

Speed: speed depends from CPU and is reverse proportional to the number of templates of the current set. To recognize and align a form, based ona  set of 10 templates, it take less than a second on a "medium" hardware (Pentium IV, 2.0Ghz).

Input: mono-chromatic images with any resolution. DIB Handle accepted. File formats supported:  TIFF, BMP, GIF, PNG, PDF.

Output: Output is made of:

 

Platform: Windows 9x/NT/Millennium/200x/XP/Vista/7 supported.

Packaging: DLL 32 bit (2 MB).

Licensing: SDK license with royalties-free runtimes. The library is royalties free till 1000 runtimes, so it can be distributed with any end-user application without paying licenses fee: if you need to distribute a larger number of runtimes you have to buy further licenses of the product.

Evaluation Version
You can download an evaluation version of this product for Visual Basic, Visual C++ or Delphi

You can also download an evaluation demo tool.

Pricing & Ordering Info
The price of Form identification and Alignment Library is EUR 5.000,00. You can order on-line in our e-commerce store or using our contacts page.

You can also buy just the Form Identification library (no alignment feature) at EUR 3.000,00. You can order on-line in our e-commerce store or using our contacts page.

More Info
If you need further info on this product, please use the contacts page.