recognizeText method

Starts text capture and recognition for a single image. Captured image will be processed according to the passed settings and recognized text will be returned with some technical information such as warnings, if any, and text orientation on the image.

export async function recognizeText(settings)

Parameters

settings
Object specifying parameters of the text capture scenario for a single image (see Options).

Return values

The method returns result depending on how the scenario finished (see Result).

Options

The table below describes parameters that you can pass as the settings argument to change text capture settings. Omitting a parameter means that a default setting will be used.

Parameter Value type Description
licenseFileName string

The name of the license file. This file must be located in the /assets/ directory in your project.

Default: "MobileCapture.License".

imageUri string

imageUri (string): Image source for the operation passed as URI. This parameter can be set to the following values:

  • "Base64": you can pass an encode base64 image string, i.e. "data:image/jpeg;base64,<base64 encoded image data>"
  • "File": image file address, i.e. "file:///data/user/0/com.abbyy.rtr.reactnative.sample.coreapi/files/pages/page_848b121d-5a7a-4ead-94e6-dd91bc6bfead.jpg".
    'content://' schema is also supported for Android.

Required parameter.

areaOfInterest object

A rectangle specifying the area of interest in the coordinates, set by the following parameters:

  • top (int): upper left point ordinate
  • bottom (int): bottom left point ordinate
  • left (int): upper left point abscissa
  • right (int): upper right point abscissa

I.e., {top: 100, bottom: 1000, left: 100, right: 1000}

 Note: All parameters should be defined to set the areaOfInterest parameter. If some of them are defined and some are not, an error occurs.

Default: a rectangle, containing the whole image.

isTextOrientationDetectionEnabled boolean

Enables or disables detection of the image orientation while preprocessing.

If the property is set to true, the image top is detected and correct orientation can be used for image rotation.

You can set this property to false for speeding the process up.

 Note: Disable the image detection only if you can be sure that the captured image has correct orientation. Otherwise the text on image will not be detected and recognized.

The default value of this property is true (enabled).

recognitionLanguages string[]

List of languages, supported for the text recognition, i.e. ['English', 'Russian'].

See the full list of supported languages here.

Default: ['English'].

Result

This section describes the object that represents text capture results. Returned parameters depend on the captured and passed image.

Parameter Value type Description
orientation int

An angle on which the image was rotated to get normal orientation. Possible values are: 0, 90, 180, 270.

warnings string[]

Warnings that occurred during processing, if any.

text string The whole recognized text as a single string. Text lines are divided with "\n". Text blocks are divided with "\n\n".
textBlocks object[] An array of text blocks. Each text block is represented as an array of textLines parameter.
textBlocks.textLines object[]

An array of text lines. Each text line has the following parameters:

  • text (string): recognized text.
  • quadrangle (object): quadrangle containing text, represented by an array of its four vertices' coordinates: { x: int, y: int }. The vertices are indexed clockwise starting from the bottom left.
  • rect (object): bounding rectangle for quadrangle. The rectangle is set by the following parameters:
      • top (int): upper left point ordinate
      • bottom (int): bottom left point ordinate
      • left (int): upper left point abscissa
      • right (int): upper right point abscissa
  • charInfo (object[]): an array, which elements contain information concerning concrete recognized symbol.
textBlocks.textLines.charInfo object[]

An array, which elements contain information concerning concrete recognized symbol. This object contains parameters, describing the symbol's place at an image, and its recognition confidence.

  • quadrangle (object): quadrangle containing text, represented by an array of its four vertices' coordinates: { x: int, y: int }
  • rect (object): bounding rectangle for quadrangle. The rectangle is set by the following parameters:
      • top (int): upper left point ordinate
      • bottom (int): bottom left point ordinate
      • left (int): upper left point abscissa
      • right (int): upper right point abscissa
  • isUncertain (boolean): returned only in case the symbol is uncertain.

Example of a result JSON.

02.03.2022 12:59:15

Usage of Cookies. In order to optimize the website functionality and improve your online experience ABBYY uses cookies. You agree to the usage of cookies when you continue using this site. Further details can be found in our Privacy Notice.