There are only two ways to collect data from forms!

The choice is simple.

Manual data entry is a boring and troublesome method. It implies a number of problems: delays in data capture process, great amount of operator�s misprints, high expenses for labor, rent of space, etc.

There is an alternative to all these problems: automatic data capture system. ABBYY FINEREADER Handprint FORMS interprets any type of forms many times faster and up to 20 times more accurately than any professional operator, efficiently helping you to collect data from paper forms.

FineReader interprets hand print, machine print, check boxes and barcodes from paper forms gathered from fax machines and scanners. After automatically processing each form, FineReader highlights the indefinite and invalid symbols for quick operator attention. Because FineReader processes all the information, entry operators spend just seconds verifying indefinitely recognized symbols rather than minutes manually keying in entire forms. In addition, FineReader can automatically check the recognized data for validity against databases, dictionaries, specific formatting, custom rules etc.

In the end, the data produced by FineReader looks exactly like the data you get from your current manual system. But you get it faster, and it is more accurate and reliably correct. Your business process is fast, accurate and effective.

Processing of all kinds of forms.

FineReader can process many kinds of forms, even those not specifically intended for machine input.

Due to various methods of separating of data and field borders FineReader copes with any kind of field marking type on the form. Now in addition to color and raster forms, black & white forms can be processed too.

With FineReader you can even process legacy (non-machine-readable) forms. This is possible due to a unique FlexiForms Technology.

Benefits

Accurate

According to the comparative tests the number of mistakes made by FineReader is several times less than the number of mistakes made by other OCR-systems. Since launch in the summer 1998 ABBYY FineReader received 19 awards from leading IT magazines worldwide for the unsurpassed accuracy of recognition. FineReader recognizes handprinted symbols making less than 5 mistakes per 1000 symbols*, this being 5 times less than a professional typist does in the morning, and 10 times less than this same typist does in the evening.

* When the form contains no corrections and is filled in accurately

Automatic control of recognition results (with crosschecks, validation rules, against a dictionary etc.) provides a completely correct and valid information.

Fast

FineReader Handprint spends 1-10 seconds per form (for A4 format), depending on form and hardware, that is more than 20 times faster than a professional typist.

Flexible

• By using FineReader API you can tune the system�s interface, you can run FineReader from other applications, and you can easily integrate FineReader into any other information processing system.
• By using FineReader API you can easily add application-specific capabilities or create custom validation rules and checks to fit your requirements.
• By using FineReader Developer Edition you can integrate FineReader into your own applications, so that you will be able to use FineReader�s functionality the way you prefer.

Scalable

FineReader can be easily scaled from one to whole number of workstations, the work can be distributed between scanning, recognizing and verifying workstations, and the number of these workstations is limited only by your needs.

Efficient

FineReader�s high form processing speed, perfect quality of recognition, automatic control of recognition results, and small cost makes it the most efficient automatic data capture system. FineReader Handprint Forms is suitable for volumes ranging from a few forms to several hundreds a minute. In all cases, the payback time is extremely short. FineReader Handprint Forms really saves your money!

Preliminary

Template designing

The only thing to do before starting the form processing is to create the form template. Designing a form template is simple and straightforward with FineReader Template Designer�s friendly interface.

You can design up to 99 different form templates for one batch. This means that FineReader is able to distinguish between that number of different forms in one pile, so you don�t have any need to sort them before processing. This means that FineReader is also able to process multi-paged forms.

The first step is to scan or import an image of a blank form. Then the system automatically finds certain form elements such as separator lines, static text blocks, checkboxes, black squares and barcodes. This eliminates the routine work of their manual selection.

After that you specify the fields that will contain text. For each of them you assign a name of a database field that corresponds to the form field.

The next step is to specify field properties, such as language, text type, block marking type, and attach dictionaries and validation rules to the fields.

Specifying validation rules, field languages, attaching dictionaries

You can skip setting any validation rules and creating specific languages for field contents if you are planning to input only several hundreds of forms. When the number of forms to be processed is much bigger, you would probably like to spend a little more time on specifying user's languages for fields and validation rules, as it improves recognition quality significantly and also drastically reduces time spent on verification.

May use a regular expression as a template for the field contents, so as to make sure that it is filled in correctly. The system also features a possibility to compare the sum in digits with the sum in writing; this feature is crucial when important financial data is processed.

If even this rich set of rules do not match your needs, you still may use Visual Basic to write your own custom rules, which will work the same way as built-in rules do.

User's languages help FineReader to interpret the information in the field correctly. For example, if you have a field containing only names, they will be recognized much more correctly if you provide FineReader with a list of names. And if you know that some fields will contain a restricted set of symbols (for example, numeric fields don�t contain letters) you can specify the set of characters that can be encountered in any particular field. FineReader Language support is a great tool to improve the recognition quality dramatically.

Scanning

FineReader supports industry standard TWAIN protocol. Its advanced support of TWAIN allows to switch off driver's own dialog boxes and control the scanner directly from FineReader. If you prefer to use any other scanning program, you just have to make sure that it saves images in one of the large set of formats supported by FineReader.

FineReader includes advanced algorithms for adaptive threshold detection, image deskew, automatic detection of image orientation, compensation for the possible non-linear distortion of image and noise removal. This set of tools is applied dynamically during scanning, image import or recognition. Now you can get good quality images for better recognition even from not so advanced scanners.

FineReader supports the following scanners: Acer, AGFA, Artec, Avigramm, Avision, Canon, Compeye, E-Lux, Epson, Fujitsu, Genius, HP, IBM, Kodak, Kye, Logitech, Lumina, Microtek, Minolta, Mitsubishi, Mustek, Nikon, Optrox, Packard Bell, PIE, Plustek, Primax, Proton, Qtronox, Relisys, Ricoh, Samsung, Scanpaq, Scanvision, Silitek, Spot, Storm, Tamarack, Teco, Ultima, UMAX, Visioneer, Xerox, and many more than 100 models of 100% TWAIN compatible scanners.

Recognition

FineReader can recognize handprinted text without tuning to any particular handwriting. It recognizes printed or typed text, automatically defining the style of typographical letters. It recognizes the most popular barcodes: EAN13, EAN8, Check Code 3, Check Interleaved 25, Code 39, CodeK128 etc. And, of course, it recognizes checkmarks of various types: checked, circled, crossed etc.

You don�t have to worry if fields are very close to each other. Advanced automatic field adjustment technology will process your form correctly.

And what if some of the forms and filled in by hand while others are typed? It is no problem for FineReader with its intelligent automatic text type detection technology.

For a relatively small number of forms you can use one computer and one copy of FineReader to accomplish the task. However, for large volume it is necessary to use several workstations devoted to specific tasks and interconnected into the network.

Recognition stations can work unattended, so collected forms may be processed day and night. You can use as many recognition stations as you need to provide necessary throughput.

Verification

Verification is the only stage at which the throughput of the whole system is affected by human productivity. That is why we paid special attention to this stage. We provided several tools and techniques to be used separately or in combination to organize efficient verification.

How to process different fields during the verification is defined at the form template designing stage. By changing template settings and adding rules you can fine-tune the verification process to minimize human efforts and solve specific tasks. The three-step verification technology implemented in FineReader (group, fragment and context verification) provides the flexibility that will help you to build the most efficient processing technology.

For example, you can include all fields containing digits in the group verification stage, and all text fields -- in fragment verification stage.

You can use rules to eliminate verification of the fields that are reliably correct. For each field you can specify how tolerant FineReader shall be in marking indefinitely recognized symbols.

Or you can use these rules to check the validity of the entered data automatically, for example, the data may be checked against a database or user-defined dictionary; you can perform date, currency and character-specific formatting validations, etc.

In addition to predefined types of rules (now there are 16 of them) you can create your own rules with VBA.

FineReader features a whole lot of other opportunities -- just start using the product and you will be constantly surprised by the capabilities it provides.

Group verification allows to find out incorrectly recognized symbols among correct ones

During the fragment verification you can see the image of the symbol and the recognition result displayed under each other. Indefinitely recognized symbols are highlighted. You don't need to strain your eyes to compare them. Also the part of the original image is displayed if the symbol context is necessary to be seen.

Export

When all the steps are done, it's time to export your data for further processing. FineReader provides you with several alternatives of export.

Into the DBF, Microsoft Excel, CSV (Comma Separated Values) file formats. Through the ODBC into any database. Through the custom export module. This is the most flexible way requiring some programming. You can use FineReader Development Tools and Visual Basic (or any programming language supporting OLE Automation) to create OLE Automation server module that will receive the recognized data. This module can then handle data any way you want.

Image export

You can also export images into databases or electronic archives in popular graphic formats: BMP, TIFF, WMF, PCX etc.

FlexiForms

If you have forms printed on the paper from different sources, and the principle fields are positioned differently on them, then you are dealing with a flexible form. The simplest example of such a form is invoice. Till now it was impossible to process such flexible forms, but now you have a software solution for this complicated task -- ABBYY FineReader Handprint Forms.

FlexiForms FineReader�s method of searching for fields is similar to one we humans use. You just have to tell FineReader once how to find all the fields on your document. The description looks like �the field is marked by word "ZIP"�. It can be the case that this word is missing, so you may want to give the application some additional tips, such as �a black line is very likely to be on the left and the text below may be something like "Street address"�.

Using a description of this kind FineReader FlexiForms searches for data on your forms. The set of such descriptions of a form is called a flexible template.

The development of a template for flexible form is a complicated task requiring intrinsic knowledge FlexiForms technology. But this is no problem, as our developers are always ready to design any template you need.

FineReader FlexiForms technology, if implemented for your forms, makes it possible to apply all the capabilities of FineReader form input, including scanning, verification, validation rules, etc.

This technology can be also applied to legacy forms, i.e. forms created not for machine-input, but which need to be processed.

Not only form processing task, but also image-indexing problem may be solved using FineReader FlexiForms, since FineReader can find the indexed field elsewhere on your images.

Examples

FlexiForms technology was applied to a great variety of documents that come from different countries of the world:

• Polish hand-written banking bills
• Russian payment drafts
• International VISA slips
• Ukrainian pay bills
• Lithinian library cards, bank transfer document used in Belgium (two families: the ones in BEF, the other ones in EUR), and many others

System Requirements:

• PC Pentium 100 or higher is recommended
• Windows 95 or Windows 98 (Russian version required to work with Russian interface) or Windows NT 4.0 (Service Pack 3 is included on the FineReader CD and will be automatically installed by FineReader setup program)
• 16 Mbytes of RAM for Windows 95 and Windows 98, 32 Mbytes of RAM for Windows NT 4.0
• 50 Mbytes of hard disk space for the minimal program installation
• 10 Mbytes of hard disk space for the program operation
• 30 Mbytes of hard disk space for the swap file
• 100% Twain-compatible scanner or fax-modem
• CD-ROM drive
• LPT port
• Mouse or other pointing device
• Color monitor