Text Recognizers

Overview of text recognizers.

Text recognizers divide an ink sample into segments and translate the ink segments into text. This is useful for recognizing handwriting. It is also useful in complex writing scenarios, such as recognizing individual kanji glyphs and offering users automatic completion of a character while they are writing it.

A text recognizer returns a Unicode string for each ink segment. Accompanying the string is the confidence level that the recognizer assigns to the result and a mapping of the result back to the original ink. This mapping shows which segment in the original ink corresponds to the Unicode code string. The string that represents the output of the text recognizer is always represented in the linear order, rather than in the spatial order or chronological order in which the segments were inputted.