find_text🔗
Short description🔗
find_text — Find text in an image.
Signature🔗
find_text( image Image, text_model TextModel, out text_result TextResultID )
Description🔗
find_text finds text specified by the TextModel in the
Image and returns the result in TextResultID.
Specific results of TextResultID can be obtained with
get_text_result and get_text_object.
The text results in TextResultID differ, depending on how
Mode was set when creating the TextModel with
create_text_model_reader.
For a text model with Mode = 'auto' find_text
extracts and classifies text of arbitrary size within the Image.
It is possible to restrict the search to characters with specific
attributes, see set_text_model_param for details.
For a text model with Mode = 'manual' reasonable
parameters for the text model, including the expected character height and
width, need to be set using set_text_model_param.
Furthermore, the Image is preprocessed such that character like
structures are enhanced. The resulting compensated image can be queried from
TextResultID if 'persistence' was set to 'true' in
TextModel.
Both the text models with Mode = 'auto' and
Mode = 'manual' apply various thresholds to the input
image and segment character candidates based on region and gray value
features. These candidates are further clustered to lines. Each line is
individually completed and tested if it fulfills the constraints of a text
line. Text models with Mode set to 'auto' require a text
line to consist of at least two characters, whereas text models with
Mode set to 'manual' require a text line to consist of at
least three characters. In a further step, punctuations and separators are
added if the correspondent parameters were set via
set_text_model_param. Finally, if TextModel contains
'text_line_structure' entries, the completed line is fitted to
these structures.
find_text finds text which is roughly horizontally aligned
in Image. text_line_orientation and rotate_image
can be used to achieve this alignment.
Since memory is allocated for the text result during the call of
find_text and during the following operations, the text result
should be
freed explicitly by the operator clear_text_result as soon as it is
no longer used.
Execution information🔗
Execution information
-
Multithreading type: reentrant (runs in parallel with non-exclusive operators).
-
Multithreading scope: global (may be called from any thread).
-
Automatically parallelized on internal data level.
This operator returns a handle. Note that the state of an instance of this handle type may be changed by specific operators even though the handle is used as an input parameter by those operators.
This operator supports canceling timeouts and interrupts.
Parameters🔗
Image (input_object) singlechannelimage → object (byte / uint2)
Input image.
TextModel (input_control) text_model → (handle)
Text model specifying the text to be segmented.
TextResultID (output_control) text_result → (handle)
Result of the segmentation.
Example🔗
(HDevelop)
read_image (Image, 'numbers_scale')
create_text_model_reader ('auto', 'Document_Rej.omc', TextModel)
* Optionally specify text properties
set_text_model_param (TextModel, 'min_char_height', 20)
find_text (Image, TextModel, TextResultID)
* Return character regions and corresponding classification results
get_text_object (Characters, TextResultID, 'all_lines')
get_text_result (TextResultID, 'class', Class)
Result🔗
If the parameters are valid, the operator
find_text returns the value 2 (H_MSG_TRUE). If
necessary, an exception is raised.
Combinations with other operators🔗
Combinations
Possible predecessors
create_text_model_reader, set_text_model_param, text_line_orientation, text_line_slant
Possible successors
Module🔗
OCR/OCV