FIELD: physics.
SUBSTANCE: method for recognizing symbols in an image includes: obtaining an image containing symbols for recognition; analyzing the resulting image to locate the area of interest; comparing the first part of the obtained image with a set of standards in the detected area of interest to establish the first value representing the accuracy of coincidence for each reference in the set of standards; creating the first set of hypotheses, at least on the basis of the established first values representing the accuracy of coincidence; comparing the second part of the resulting image with a set of the selected standards to establish the second value indicating the accuracy of coincidence for each standard in the set of the selected standards. A set of the selected standards is formed on the basis of the first set of hypotheses; creating the second set of hypotheses, at least based on the established second values that represent the accuracy of coincidence; and recognizing the symbols in the resulting image, at least based on the first set of hypotheses and the second set of hypotheses.
EFFECT: ensuring the recognition of symbols not completely represented in the image.
27 cl, 8 dwg
| Title | Year | Author | Number |
|---|---|---|---|
| TRACELESS IMAGE CAPTURE, USING MOBILE DEVICE | 2020 |
|
RU2787136C2 |
| METHOD OF DETECTING NECESSITY OF STANDARD LEARNING FOR VERIFICATION OF RECOGNIZED TEXT | 2014 |
|
RU2641225C2 |
| COMPARING DOCUMENTS USING RELIABLE SOURCE | 2014 |
|
RU2597163C2 |
| HANDWRITING RECOGNITION USING NEURAL NETWORKS | 2020 |
|
RU2757713C1 |
| METHOD AND SYSTEM FOR VERIFICATION DURING READING | 2014 |
|
RU2571396C2 |
| APPARATUS AND METHOD OF SEARCHING FOR DIFFERENCES IN DOCUMENTS | 2013 |
|
RU2571378C2 |
| NEURAL NETWORK TRAINING BY MEANS OF SPECIALIZED LOSS FUNCTIONS | 2018 |
|
RU2707147C1 |
| VIDEO CAPTURE IN DATA INPUT SCENARIO | 2015 |
|
RU2613734C1 |
| OPTICAL CHARACTER RECOGNITION USING SPECIALIZED CONFIDENCE FUNCTIONS, IMPLEMENTED ON THE BASIS OF NEURAL NETWORKS | 2018 |
|
RU2703270C1 |
| FUZZY SEARCH USING WORD FORMS FOR WORKING WITH BIG DATA | 2021 |
|
RU2768233C1 |
