do_ocr_single_class_knnT_do_ocr_single_class_knnDoOcrSingleClassKnnDoOcrSingleClassKnn (Operator)

Name

do_ocr_single_class_knnT_do_ocr_single_class_knnDoOcrSingleClassKnnDoOcrSingleClassKnn — Classify a single character with an OCR classifier.

Signature

do_ocr_single_class_knn(Character, Image : : OCRHandle, NumClasses, NumNeighbors : Class, Confidence)

Description

do_ocr_single_class_knndo_ocr_single_class_knnDoOcrSingleClassKnnDoOcrSingleClassKnnDoOcrSingleClassKnn uses the OCR-k-NN classifier OCRHandleOCRHandleOCRHandleOCRHandleOCRHandle to determine the best training samples for the features of the character given by the region CharacterCharacterCharacterCharactercharacter and the gray values ImageImageImageImageimage and returns their classes in ClassClassClassClassclassVal and the corresponding confidences in ConfidenceConfidenceConfidenceConfidenceconfidence. If the first NumNeighborsNumNeighborsNumNeighborsNumNeighborsnumNeighbors training vectors have different classes, maximally NumClassesNumClassesNumClassesNumClassesnumClasses classes are returned sorted by frequency and weighted distance. The confidences lie between 0.0 and 1.0. The larger the value, the more reliable is the classification. The confidence can be only computed robustly if NumNeighborsNumNeighborsNumNeighborsNumNeighborsnumNeighbors is large enough.

Because multiple classes may be returned by do_ocr_single_class_knndo_ocr_single_class_knnDoOcrSingleClassKnnDoOcrSingleClassKnnDoOcrSingleClassKnn, CharacterCharacterCharacterCharactercharacter may only contain a single region (a single character). If multiple characters should be classified in a single call, do_ocr_multi_class_knndo_ocr_multi_class_knnDoOcrMultiClassKnnDoOcrMultiClassKnnDoOcrMultiClassKnn must be used. Before calling do_ocr_single_class_knndo_ocr_single_class_knnDoOcrSingleClassKnnDoOcrSingleClassKnnDoOcrSingleClassKnn, the classifier must be trained with trainf_ocr_class_knntrainf_ocr_class_knnTrainfOcrClassKnnTrainfOcrClassKnnTrainfOcrClassKnn or select_feature_set_trainf_knnselect_feature_set_trainf_knnSelectFeatureSetTrainfKnnSelectFeatureSetTrainfKnnSelectFeatureSetTrainfKnn. If all NumNeighborsNumNeighborsNumNeighborsNumNeighborsnumNeighbors neighbors are of the same class, only one class is returned.

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

CharacterCharacterCharacterCharactercharacter (input_object) region → object

Character to be recognized.

ImageImageImageImageimage (input_object) singlechannelimage → object (byte / uint2)

Gray values of the character.

OCRHandleOCRHandleOCRHandleOCRHandleOCRHandle (input_control) ocr_knn → (handle)

Handle of the k-NN classifier.

NumClassesNumClassesNumClassesNumClassesnumClasses (input_control) integer-array → (integer)

Number of maximal classes to determine.

Default value: 1

Suggested values: 1, 2, 3, 4, 5

NumNeighborsNumNeighborsNumNeighborsNumNeighborsnumNeighbors (input_control) integer-array → (integer)

Number of neighbors to consider.

Default value: 1

Suggested values: 1, 2, 3, 4, 5

ClassClassClassClassclassVal (output_control) string(-array) → (string)

Results of classifying the character with the k-NN.

ConfidenceConfidenceConfidenceConfidenceconfidence (output_control) real(-array) → (real)

Confidence(s) of the class(es) of the character.

Result

If the parameters are valid, the operator do_ocr_single_class_knndo_ocr_single_class_knnDoOcrSingleClassKnnDoOcrSingleClassKnnDoOcrSingleClassKnn returns the value 2 (H_MSG_TRUE). If necessary, an exception is raised.