FIELD: physics.
SUBSTANCE: method comprises the step of obtaining a digital image of the document by the user electronic device. Further, according to the method, the qualification parameter of the digital image for the OCR is determined using the classifier on the user's electronic device, which is a sign of, whether the digital image is suitable for producing an OCR result of acceptable quality. Herewith the classifier is trained based on, at least, the predicted image noisiness parameter to determine the document suitability parameter for the OCR and/or the OCR probability indicator, and passes validation after learning the said classifier.
EFFECT: expanding the arsenal of technical means for analyzing the suitability of digital images.
28 cl, 5 dwg
Authors
Dates
2017-10-24—Published
2016-12-06—Filed