get_prep_info_ocr_class_mlp🔗
Short description🔗
get_prep_info_ocr_class_mlp — Compute the information content of the preprocessed feature vectors
of an OCR classifier.
Signature🔗
get_prep_info_ocr_class_mlp( ocr_mlp OCRHandle, filename.read TrainingFile, string Preprocessing, out real InformationCont, out real CumInformationCont )
Description🔗
get_prep_info_ocr_class_mlp computes the information content
of the training vectors that have been transformed with the
preprocessing given by Preprocessing.
Preprocessing can be set to 'principal_components'
or 'canonical_variates'. The OCR classifier
OCRHandle must have been created with
create_ocr_class_mlp. The preprocessing methods are
described with create_class_mlp. The information content is
derived from the variations of the transformed components of the
feature vector, i.e., it is computed solely based on the training
data, independent of any error rate on the training data. The
information content is computed for all relevant components of the
transformed feature vectors (NumInput for
'principal_components' and min(NumOutput - 1,
NumInput) for 'canonical_variates', see
create_class_mlp), and is returned in
InformationCont as a number between 0 and 1. To convert
the information content into a percentage, it simply needs to be
multiplied by 100. The cumulative information content of the first
\(n\) components is returned in the \(n\)-th component of
CumInformationCont, i.e., CumInformationCont
contains the sums of the first \(n\) elements of
InformationCont. To use
get_prep_info_ocr_class_mlp, a sufficient number of samples
must be stored in the training files given by TrainingFile
(see write_ocr_trainf).
InformationCont and CumInformationCont can be used
to decide how many components of the transformed feature vectors
contain relevant information. An often used criterion is to require
that the transformed data must represent x% (e.g., 90%) of the
total data. This can be decided easily from the first value of
CumInformationCont that lies above x%. The number thus
obtained can be used as the value for NumComponents in a
new call to create_ocr_class_mlp. The call to
get_prep_info_ocr_class_mlp already requires the creation of
a classifier, and hence the setting of NumComponents in
create_ocr_class_mlp to an initial value. However, if
get_prep_info_ocr_class_mlp is called it is typically not
known how many components are relevant, and hence how to set
NumComponents in this call. Therefore, the following
two-step approach should typically be used to select
NumComponents: In a first step, a classifier with the
maximum number for NumComponents is created
(NumInput for 'principal_components' and
min(NumOutput - 1,
NumInput) for 'canonical_variates'). Then, the
training samples are saved in a training file using
write_ocr_trainf. Subsequently,
get_prep_info_ocr_class_mlp is used to determine the
information content of the components, and with this
NumComponents. After this, a new classifier with the
desired number of components is created, and the classifier is
trained with trainf_ocr_class_mlp.
Execution information🔗
Execution information
-
Multithreading type: reentrant (runs in parallel with non-exclusive operators).
-
Multithreading scope: global (may be called from any thread).
-
Processed without parallelization.
Parameters🔗
OCRHandle (input_control) ocr_mlp → (handle)
Handle of the OCR classifier.
TrainingFile (input_control) filename.read(-array) → (string)
Names of the training files.
Default: 'ocr.trf'
File extension: .trf, .otr
Preprocessing (input_control) string → (string)
Type of preprocessing used to transform the feature vectors.
Default: 'principal_components'
List of values: 'canonical_variates', 'principal_components'
InformationCont (output_control) real-array → (real)
Relative information content of the transformed feature vectors.
CumInformationCont (output_control) real-array → (real)
Cumulative information content of the transformed feature vectors.
Example🔗
(HDevelop)
* Create the initial OCR classifier.
read_ocr_trainf_names ('ocr.trf', CharacterNames, CharacterCount)
create_ocr_class_mlp (8, 10, 'constant', 'default', CharacterNames, 80, \
'canonical_variates', |CharacterNames|, 42, OCRHandle)
* Get the information content of the transformed feature vectors.
get_prep_info_ocr_class_mlp (OCRHandle, 'ocr.trf', 'canonical_variates', \
InformationCont, CumInformationCont)
* Determine the number of transformed components.
* NumComp = [...]
* Create the final OCR classifier.
create_ocr_class_mlp (8, 10, 'constant', 'default', CharacterNames, 80, \
'canonical_variates', NumComp, 42, OCRHandle)
* Train the final classifier.
trainf_ocr_class_mlp (OCRHandle, 'ocr.trf', 100, 1, 0.01, Error, ErrorLog)
write_ocr_class_mlp (OCRHandle, 'ocr.omc')
Result🔗
If the parameters are valid, the operator
get_prep_info_ocr_class_mlp returns the value 2 (H_MSG_TRUE). If
necessary, an exception is raised.
get_prep_info_ocr_class_mlp may return the error 9211
(Matrix is not positive definite) if Preprocessing \(=\)
'canonical_variates' is used. This typically indicates
that not enough training samples have been stored for each class.
Combinations with other operators🔗
Combinations
Possible predecessors
create_ocr_class_mlp, write_ocr_trainf, append_ocr_trainf, write_ocr_trainf_image
Possible successors
Module🔗
OCR/OCV