An Overview of Recognition Modules

OCR Plus

Automatic recognition module

If the automatic recognition module is used, the Engine will try to automatically select the most suitable recognition module for the zone. This is determined just before recognition, according to the zone's filling method and, if necessary, other settings, most typically the Character Set.

MTX (Mtext) omnifont recognition module

The MTX module recognizes machine printed text from printed publications, laser or ink-jet printers, and electric typewriters. Output from mechanical typewriters in good condition, and from draft-quality, letter-quality, or near-letter quality dot-matrix printers is also acceptable.

The fastest of the selectable OCR engines
Supports up to 64 zones on one image
Supports Omnifont, Draftdot9 and Draftdot24 filling methods.
The OmniFont recognition module

detects and transmits bold, italic or underlined text (or any combination)
detects and transmits character size
classifies font types into three broad categories: serif, sans serif and monospaced.

Provides 2 page-level accuracy and speed trade off settings including a combined Accurate & Balanced value and Fast (RECGMD_FAST is respected, while RECGMD_BALANCED and RECGMD_ACCURATE are merged to one value)
Provides Checking Subsystem based correction
Handles A3 size (11.69" x 16.54") portrait and landscape images with 300 dpi resolution. (At larger resolution the supported image size is smaller.)
Supports images with the following resolution ranges: 90-110, 160-240, 280-320, 400, 600. (Does not process images larger than 6600 pixels in width or height.)
Supports character sets for the following languages, singly or in combination: English, Brazilian, Danish, Dutch, Finnish, French, German, Italian, Norwegian, Portuguese, Spanish, Swedish.

MOR (Multi-lingual Omnifont Recognition) module

The MOR module recognizes machine printed text from printed publications, laser or ink-jet printers, and electric typewriters. Output from mechanical typewriters in good condition, and from letter- or near-letter quality (LQ, NLQ) dot-matrix printers is also acceptable.

Supports up to 500 zones on one image
Supports Omnifont, Draftdot24 and OCR-A filling methods
The OmniFont recognition module

detects and transmits bold, italic or underlined text (or any combination)
detects and transmits character size
classifies font types into three broad categories: serif, sans serif and monospaced
uses contour analysis along with pattern matching to perform recognition

Provides 3 page-level accuracy and speed trade off settings including Accurate, Balanced and Fast (RECGMD_FAST is respected, while RECGMD_BALANCED and RECGMD_ACCURATE are merged to one value)
Provides Checking Subsystem-based correction
Handles A3 paper size (11.69" x 16.54") portrait and landscape images with 300 dpi resolution
Recognizes about 500 characters, including Latin, Greek and Cyrillic alphabets with enough accented letters to recognize the 119 Languages supported by the engine.

There are two ways of modifying incoming images to make them more suitable for this omnifont recognition module:

Standard mode Fax output. (200 x 100 dpi). This switch doubles the pixels in the image’s vertical direction. Faxes sent in Fine Mode (200 x 200 dpi) are preferable. However, if faxes are sent in Fine Mode, then the Standard mode Fax Output should not be used.
Draft 24-pin dot-matrix output. Use the FILL_DRAFTDOT24 filling method to have the character contours smoothed. Again, NLQ or LQ quality output can usually be better recognized without using FILL_DRAFTDOT24.

DOT Matrix (DOT 9-pin draft dot-matrix recognition module)

The Dot Matrix module is designed for ONLY draft-quality 9-pin dot-matrix texts.

For NLQ or LQ texts, the RECOGMODULE_OMNIFONT_PLUS2W, RECOGMODULE_OMNIFONT_PLUS3W, RECOGMODULE_MTEXT_OMNIFONT or RECOGMODULE_MULTI_LINGUAL_OMNIFONT modules are likely to give better results.

If FILL_DRAFTDOT9 filling method is set together with RECOGMODULE_AUTO, RECOGMODULE_MTEXT_OMNIFONT is used, provided that all characters (or languages or filters) validated for the zone are supported by it. If any are not supported, this module is used.

Supports 76 languages, of which 14 have dictionary support: Catalan, Danish, Dutch, English, Finnish, French, German, Greek, Hungarian, Italian, Norwegian, Portuguese, Spanish and Swedish.
Reads multiple languages.
Reads 18 of the 29 punctuation characters. (The Low Double Comma Quotation Mark is missing).
Supports 24 of the 55 miscellaneous characters. (Missing charcters include the Euro Sign, the Small Script F, the Copyright Sign, Registered Trade Mark Sign and the Degree Sign.)
Does not interpret the recognition trade-off setting and cannot be trained.
Used if it is directly specified in a zone structure.
Generates confidence data on recognized characters and can interpret all filter values.

OMR optical mark recognition module

For more information see the LEAD OMR Overview.

ICR-HNR hand-printed numeral recognition module

The ICR-HNR module can be used for recognition of hand-printed numerals and four additional signs. If further hand-printed characters are to be recognized, the use of the RECOGMODULE_RER_PRINTED recognition module is recommended.

This recognition module can recognize the following hand-printed characters:

Numerical Digits 0–9
Plus Sign (+)
Minus Sign (–)
Period or Full-stop (.)
Comma (,)

The filter ZONE_CHAR_FILTER_DIGIT can be used to exclude the last four characters. The filters ZONE_CHAR_FILTER_PUNCTUATION and ZONE_CHAR_FILTER_MISCELLANEOUS function, other filters have no effect.

ICR-RER hand-printed recognition module

The ICR-RER module is a third-party recognition module from reRecognition GmbH, Germany. The Engine contains its recognition engine version 4.2f.

This recognition module can be used for recognition of hand-printed alphanumerical characters, i.e. upper and lower case letters, the digits and some others. Although it can be used to read flowing text, its main application area is in form-like situations, where the form designer has great control over the content and maybe length of hand-printed information given in each zone.

MAT matrix matching recognition module

The MAT module is designed to read certain groups of fixed-font characters specially designed for OCR or imaging applications, in which no two characters have similar shapes. Each character group has its own filling method. Application areas are in banking, check or waybill handling, product distribution and document validation, where high accuracy can be vital. It also handles some non-fixed print styles.

FILL_OCRA

OCR-A. Uppercase English letters (26), digits, some punctuation, and 3 special OCR-A symbols:

(OCR Chair)
(OCR Hook)
(OCR Fork:)

FILL_OCRB

OCR-B. Uppercase English letters (26), digits and some punctuation.

FILL_MICR

MICR (E-13B). Digits plus some punctuation and 4 special MICR symbols:

(OCR Branch Bank)
(OCR Amount of Check)
(OCR Dash)
(OCR Customer Account Number)

FILL_DOTDIGIT

Ten digits only and the period. Commas are also read, but converted to periods. Though this is in theory a fixed-font, many variants of it are used.

FILL_DASHDIGIT

Ten digits only and the period. Commas are also read, but converted to periods. Though this is in theory a fixed-font, many variants of it are used.

FRX (FireWorX) multi-lingual omnifont recognition module

The FRX module recognizes machine printed text from printed publications, laser or ink-jet printers, and electric typewriters. Output from mechanical typewriters in good condition, and from letter- or near-letter quality (LQ, NLQ) dot-matrix printers is also acceptable.

Optimized for speed
Supports up to 2,500 zones on one image
Supports Omnifont filling methods
Supports Latin, Greek and Cyrillic alphabets with enough accented letters to recognize the 54 languages (Languages and modules)

PLUS2W and PLUS3W omnifont recognition modules

The PLUS modules recognize machine printed text from printed publications, laser or ink-jet printers and electric typewriters. Output from mechanical typewriters in good condition may also be acceptable.

The PLUS2W and PLUS3W modules use voting technology to provide improved recognition results.
The PLUS2W module combines the results from the MOR and MTX modules.
The PLUS3W module combines the results from the MOR, MTX and FRX modules.

With either of these two voting modules, the accuracy is considerably better, but the recognition may need significantly more time than any single module.

Note: The Plus engine has global and module specific limits on image size. The global (meaning on loading) limits are 8400 pixels and 22 inches. Some of the OCR modules have even lower limits. The most limiting modules are MOR and MTX. Both have a hard coded upper limit. The limit is 5440 pixels in the case of MOR and 6688 for MTX. MTX has an additional limit based on the image area. It is limited to 20 megapixels (in practice this means something near 4000x5000 pixels).

OCR Professional

Automatic recognition module

If the automatic recognition module is used, the engine will try to automatically select the most suitable recognition module for the zone. This is determined just before recognition, according to the zone's filling method and, if necessary, other settings, most typically the Character Set.

MTX (Mtext) omnifont recognition module

Supports Omnifont filling methods.
The OmniFont recognition module
- Detects and transmits bold, italic or underlined text (or any combination)
- Detects and transmits character size
- Classifies font types into three broad categories: serif, sans serif and monospaced.

Provides two page-level accuracy and speed trade off settings including a combined Accurate & Balanced value and Fast (RECGMD_FAST is respected, while RECGMD_BALANCED and RECGMD_ACCURATE are merged to one value)
Supports character sets for the following languages, singly or in combination: English, Brazilian, Danish, Dutch, Finnish, French, German, Italian, Norwegian, Portuguese, Spanish, Swedish.

MOR (Multi-lingual Omnifont Recognition) module

Supports up to 500 zones on one image.
Supports Omnifont and Draftdot24 filling methods.
The OmniFont recognition module
- Detects and transmits bold, italic or underlined text (or any combination)
- Detects and transmits character size
- Classifies font types into three broad categories: serif, sans serif and monospaced
- Contour analysis-based recognition can be supplemented this with an innovative form of pattern matching not requiring enormous pre-stored shape libraries

Provides 3 page-level accuracy and speed trade off settings including Accurate, Balanced and Fast (RECGMD_FAST is respected, while RECGMD_BALANCED and RECGMD_ACCURATE are merged to one value)
Recognizes about 500 characters, including Latin, Greek and Cyrillic alphabets with enough accented letters to recognize the 119 Languages supported by the engine.
There are two ways of modifying incoming images to make them more suitable for this omnifont recognition module:

Standard mode Fax output. (200 x 100 dpi). This switch doubles the pixels in the image’s vertical direction. Faxes sent in Fine Mode (200 x 200 dpi) are preferable. However, if faxes are sent in Fine Mode, then the Standard mode Fax Output should not be used.

Draft 24-pin dot-matrix output. Use the FILL_DRAFTDOT24 filling method to have the character contours smoothed. Again, NLQ or LQ quality output can usually be better recognized without using FILL_DRAFTDOT24.

DOT Matrix (DOT 9-pin draft dot-matrix recognition module)

The Dot Matrix module is designed for ONLY draft-quality 9-pin dot-matrix texts.

For NLQ or LQ texts, the RECOGMODULE_OMNIFONT_PLUS2W, RECOGMODULE_OMNIFONT_PLUS3W, RECOGMODULE_MTEXT_OMNIFONT or RECOGMODULE_MULTI_LINGUAL_OMNIFONT modules are likely to give better results.

Supports 76 languages, of which 14 have dictionary support: Catalan, Danish, Dutch, English, Finnish, French, German, Greek, Hungarian, Italian, Norwegian, Portuguese, Spanish and Swedish.
Reads multiple languages.
Reads 18 of the 29 punctuation characters. (The Low Double Comma Quotation Mark is missing).
Supports 24 of the 55 miscellaneous characters. (Missing charcters include the Euro Sign, the Small Script F, the Copyright Sign, Registered Trade Mark Sign and the Degree Sign.)
Does not interpret the recognition trade-off setting and cannot be trained.
Used if it is directly specified in a zone structure.
Generates confidence data on recognized characters and can interpret all filter values.

OMR optical mark recognition module

For more information see the LEAD OMR Overview.

ICR-HNR hand-printed numeral recognition module

This ICR-HNR module can be used for recognition of hand-printed numerals and four additional signs. If further hand-printed characters are to be recognized, the use of the DOC2_RECOGMODULE_RER_PRINTED recognition module is recommended.

This recognition module can recognize the following hand-printed characters:

The digits (0-9)
The Plus Sign (+)
The Minus Sign (–)
The Period or Full-stop (.)
The Comma (,).

The filter DOC2_ZONE_CHAR_FILTER_DIGIT can be used to exclude the last four characters. The filters DOC2_ZONE_CHAR_FILTER_PUNCTUATION and DOC2_ZONE_CHAR_FILTER_MISCELLANEOUS function, other filters have no effect.

ICR-RER hand-printed recognition module

The ICR-RER module is a third-party recognition module from reRecognition GmbH, Germany. The engine contains its recognition engine version 4.2f.

MAT matrix matching recognition module

DOC2_FILL_OCRA

OCR-A. Uppercase English letters (26), digits, some punctuation and 3 special OCR-A symbols:

$image\ukn_OCRChair.gif$ (OCR Chair)

$image\ukn_OCRHook.gif$ (OCR Hook)

$image\ukn_OCRHook.gif$ (OCR OCRFork)

DOC2_FILL_OCRB

OCR-B. Uppercase English letters (26), digits and some punctuation.

DOC2_FILL_MICR

MICR (E-13B). Digits plus some punctuation and 4 special MICR symbols:

$image\ukn_OCRBranch.gif$ (OCR Branch Bank)

$image\ukn_OCRAmount.gif$ (OCR Amount of Check)

$image\ukn_OCRDash.gif$ (OCR Dash)

$image\ukn_OCRAccount.gif$ (OCR Customer Account Number)

DOC2_FILL_DOTDIGIT

Ten digits only and the period. Commas are also read, but converted to periods. Though this is in theory a fixed-font, many variants of it are used.

DOC2_FILL_DASHDIGIT

Ten digits only and the period. Commas are also read, but converted to periods. Though this is in theory a fixed-font, many variants of it are used.

FRX (FireWorX) multi-lingual omnifont recognition module

Supports Latin, Greek and Cyrillic alphabets with enough accented letters to recognize the 54 languages (Languages and modules).

PLUS2W and PLUS3W omnifont recognition modules

The PLUS2W and PLUS3W modules use voting technology to provide improved recognition results.
The PLUS2W module combines the results from the MOR and MTX modules.
The PLUS3W module combines the results from the MOR, MTX and FRX modules.

With any of these two voting modules, the accuracy is considerably better, but the recognition may need significantly more time than any single module.

Note: For the Professional engine the image handling algorithms in all modules were changed to deal with practically unlimited image sizes: 32000 pixels in both directions

The following table shows the text recognition module support for each of the 119 languages (General for both Plus and Professional engines):

Language	MOR	MTX	FRX	PLUS2W	PLUS3W	DOT	RER
Afrikaans	Yes	No	Yes	Yes	Yes	Yes	Yes
Albanian	Yes	No	Yes	Yes	Yes	Yes	Yes
Aymara	Yes	No	Yes	Yes	Yes	Yes	Yes
Basque	Yes	No	Yes	Yes	Yes	Yes	Yes
Bemba	Yes	Yes	No	Yes	Yes	Yes	Yes
Blackfoot	Yes	Yes	No	Yes	Yes	Yes	Yes
Brazilian	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Breton	Yes	No	Yes	Yes	Yes	Yes	Yes
Bugotu	Yes	Yes	No	Yes	Yes	Yes	Yes
Bulgarian	Yes	No	Yes	Yes	Yes	No	No
Byelorussian	Yes	No	Yes	Yes	Yes	No	No
Catalan	Yes	No	Yes	Yes	Yes	Yes	Yes
Chamorro	Yes	No	No	Yes	Yes	Yes	Yes
Chechen	Yes	No	No	Yes	Yes	Yes	Yes
Corsican	Yes	No	No	Yes	Yes	Yes	Yes
Croatian	Yes	No	Yes	Yes	Yes	No	Yes
Crow	Yes	Yes	No	Yes	Yes	Yes	Yes
Czech	Yes	Yes	No	Yes	Yes	No	Yes
Danish	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Dutch	Yes	Yes	Yes	Yes	Yes	Yes	Yes
English	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Eskimo (Inuit)	Yes	No	Yes	Yes	Yes	No	Yes
Esperanto	Yes	No	No	Yes	Yes	No	No
Estonian	Yes	Yes	No	Yes	Yes	Yes	Yes
Faroese	Yes	No	Yes	Yes	Yes	No	No
Fijian	Yes	No	No	Yes	Yes	No	Yes
Finnish	Yes	Yes	Yes	Yes	Yes	Yes	Yes
French	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Frisian	Yes	No	Yes	Yes	Yes	Yes	Yes
Friulian	Yes	No	Yes	Yes	Yes	Yes	Yes
Gaelic (Irish)	Yes	No	Yes	Yes	Yes	Yes	Yes
Gaelic (Scottish)	Yes	No	Yes	Yes	Yes	Yes	Yes
Galician	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Ganda	Yes	No	No	Yes	Yes	No	Yes
German	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Greek	Yes	No	Yes	Yes	Yes	Yes	No
Guarani	Yes	No	No	Yes	Yes	Yes	Yes
Hani	Yes	Yes	No	Yes	Yes	Yes	Yes
Hawaiian	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Hungarian	Yes	No	Yes	Yes	Yes	Yes	Yes
Icelandic	Yes	No	Yes	Yes	Yes	No	No
Ido	Yes	Yes	No	Yes	Yes	Yes	Yes
Indonesian	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Interlingua	Yes	Yes	No	Yes	Yes	Yes	Yes
Italian	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Kabardian	Yes	No	No	Yes	Yes	No	No
Kasub	Yes	No	No	Yes	Yes	No	Yes
Kawa	Yes	Yes	No	Yes	Yes	Yes	Yes
Kikuyu	Yes	No	No	Yes	Yes	No	No
Kongo	Yes	Yes	No	Yes	Yes	Yes	Yes
Kpelle	Yes	Yes	No	Yes	Yes	Yes	Yes
Kurdish	Yes	No	Yes	Yes	Yes	No	Yes
Latin	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Latvian	Yes	No	Yes	Yes	Yes	No	Yes
Lithuanian	Yes	No	Yes	Yes	Yes	No	Yes
Luba	Yes	No	No	Yes	Yes	No	Yes
Luxembourgian	Yes	No	No	Yes	Yes	Yes	Yes
Macedonian	Yes	No	Yes	Yes	Yes	No	No
Malagasy	Yes	Yes	No	Yes	Yes	Yes	Yes
Malay	Yes	No	Yes	Yes	Yes	No	Yes
Malinke	Yes	No	No	Yes	Yes	Yes	Yes
Maltese	Yes	No	No	Yes	Yes	No	No
Maori	Yes	Yes	No	Yes	Yes	Yes	Yes
Mayan	Yes	No	No	Yes	Yes	Yes	Yes
Miao	Yes	Yes	No	Yes	Yes	Yes	Yes
Minankabaw	Yes	No	No	Yes	Yes	No	Yes
Mohawk	Yes	Yes	No	Yes	Yes	Yes	Yes
Moldavian	Yes	No	No	Yes	Yes	No	No
Nahuatl	Yes	Yes	No	Yes	Yes	Yes	Yes
Norwegian	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Nyanja	Yes	Yes	No	Yes	Yes	Yes	Yes
Occidental	Yes	No	No	Yes	Yes	Yes	Yes
Ojibway	Yes	No	No	Yes	Yes	No	Yes
Papiamento	Yes	No	No	Yes	Yes	Yes	Yes
Pigin English	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Polish	Yes	No	Yes	Yes	Yes	No	Yes
Portuguese	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Provençal	Yes	No	No	Yes	Yes	Yes	Yes
Quechua	Yes	No	No	Yes	Yes	Yes	Yes
Rhaetic	Yes	No	No	Yes	Yes	Yes	Yes
Romanian	Yes	No	Yes	Yes	Yes	No	No
Romany	Yes	No	No	Yes	Yes	No	Yes
Ruanda	Yes	Yes	No	Yes	Yes	Yes	Yes
Rundi	Yes	Yes	No	Yes	Yes	Yes	Yes
Russian	Yes	No	Yes	Yes	Yes	No	No
Sami	Yes	No	No	Yes	Yes	No	Yes
Sami, Lule	Yes	No	No	Yes	Yes	No	Yes
Sami, Northern	Yes	No	No	Yes	Yes	No	Yes
Sami, Southern	Yes	No	No	Yes	Yes	No	Yes
Samoan	Yes	No	No	Yes	Yes	Yes	Yes
Sardinian	Yes	No	No	Yes	Yes	Yes	Yes
Serbian	Yes	No	Yes	Yes	Yes	No	No
Serbian, Latinic	Yes	No	Yes	Yes	Yes	No	Yes
Shona	Yes	Yes	No	Yes	Yes	Yes	Yes
Sioux	Yes	Yes	No	Yes	Yes	Yes	Yes
Slovak	Yes	No	Yes	Yes	Yes	No	Yes
Slovenian	Yes	No	Yes	Yes	Yes	No	Yes
Somali	Yes	Yes	No	Yes	Yes	Yes	Yes
Sorbian (Wend)	Yes	No	Yes	Yes	Yes	no	Yes
Sotho	Yes	No	No	Yes	Yes	Yes	Yes
Spanish	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sundanese	Yes	No	No	Yes	Yes	Yes	Yes
Swahili	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Swazi	Yes	No	No	Yes	Yes	No	Yes
Swedish	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Tagalog	Yes	Yes	No	Yes	Yes	Yes	Yes
Tahitian	Yes	No	Yes	Yes	Yes	Yes	Yes
Tinpo	Yes	Yes	No	Yes	Yes	Yes	Yes
Tongan	Yes	Yes	No	Yes	Yes	Yes	Yes
Tswana (Chuana)	Yes	No	No	Yes	Yes	Yes	Yes
Tun	Yes	Yes	No	Yes	Yes	Yes	Yes
Turkish	Yes	No	Yes	Yes	Yes	No	Yes
Ukrainian	Yes	No	Yes	Yes	Yes	No	No
Visayan	Yes	Yes	No	Yes	Yes	Yes	Yes
Welsh	Yes	No	Yes	Yes	Yes	Yes	Yes
Wolof	Yes	No	No	Yes	Yes	Yes	Yes
Xhosa	Yes	Yes	No	Yes	Yes	Yes	Yes
Zapotec	Yes	Yes	No	Yes	Yes	Yes	Yes
Zulu	Yes	No	Yes	Yes	Yes	No	Yes