Compute the features of a character.
get_features_ocr_class_mlp computes the features of the character given by Character with the OCR classifier OCRHandle and returns them in Features. In contrast to do_ocr_single_class_mlp and do_ocr_multi_class_mlp, the character is passed as a single image object. Hence, before calling get_features_ocr_class_mlp, reduce_domain must typically be called. The parameter Transform determines whether the feature transformation specified with Preprocessing in create_ocr_class_mlp for the classifier should be applied (Transform = 'true') or whether the untransformed features should be returned (Transform = 'false'). get_features_ocr_class_mlp can be used to inspect the features that are used for the classification.
|
Character (input_object) |
image -> object : byte / uint2 |
| Input character. | |
|
OCRHandle (input_control) |
ocr_mlp -> integer |
| Handle of the OCR classifier. | |
|
Transform (input_control) |
string -> string |
| Should the feature vector be transformed with the preprocessing? | |
| Default value: 'true' | |
| List of values: 'true', 'false' | |
|
Features (output_control) |
real-array -> real |
| Feature vector of the character. | |
If the parameters are valid, the operator get_features_ocr_class_mlp returns the value 2 (H_MSG_TRUE). If necessary an exception handling is raised.
get_features_ocr_class_mlp is reentrant and processed without parallelization.
OCR/OCV
