Countries and languages for RecoStar recognition

Language may vary from document to document or from paragraph to paragraph within a document, and even vary within a text line as single words may be written in a different language than the main text.

The country-specific style and conventions of handwriting, the general format of addresses, the most common vocabulary words, with regards to country-related spelling, are very important when configuring a recognition engine.

Also, each country may require more than one language. For example, Germany requires only German, whereas Switzerland requires German, French, Italian, and Rhaeto-Romantic.

To ensure your recognition results for multi-lingual documents are sufficient, you can either use only a country selection, or combine country and language selections.

Tip For the best results, ensure your selection as small as possible by specifying a single country instead of a country group or several languages. This is because with each additional country or language combination, the "Machine print" recognition gets slower.

For regional handwriting styles, standard lettering fonts, or a content type different from Alphanumeric, always select a single country only, and no language.

When using language groups, always limit your selection to one of the following:

  1. Cyrillic (according to Windows code page 1251)

  2. Greek (according to Windows code page 1253)

  3. Latin (according to Windows code pages 1250 for Central Europe, 1252 for Western Europe, 1254 for Turkey, and 1257 for Baltic.)

  4. Azerbaijanian