FIELD: information technology.
SUBSTANCE: invention relates to method and system for text detecting in graphic images. In the method of text detecting an image is received, and a set of bound components in the image is determined. For each bound component in the set a limiting area is determined. Set of image areas is determined based on the limiting area. Each area in the set of areas is classified and normalized based on the classification. Normalized areas set is combined into a binary image. In the combined normalized areas set a text is detected and recognized.
EFFECT: technical result is optimizing text detection in GUI screen images, in particular, improvement of accuracy and efficiency of text detection and recognition in GUI screen images.
20 cl, 14 dwg
| Title | Year | Author | Number |
|---|---|---|---|
| METHOD AND SYSTEM FOR CONVERTING SCREENSHOT INTO METAFILE | 2013 |
|
RU2534005C2 |
| METHOD FOR RECOGNITION OF TEXT IN IMAGES OF DOCUMENTS | 2021 |
|
RU2768544C1 |
| METHOD AND SYSTEM OF TEXT IMPROVEMENT AT DIGITAL COPYING OF PRINTED DOCUMENTS | 2012 |
|
RU2520407C1 |
| METHOD AND DEVICE FOR DETERMINING DOCUMENT SUITABILITY FOR OPTICAL CHARACTER RECOGNITION (OCR) | 2016 |
|
RU2634195C1 |
| IMPROVED CONTRAST AND NOISE REDUCTION ON IMAGES OBTAINED FROM CAMERAS | 2017 |
|
RU2721188C2 |
| METHOD OF EDITING STATIC DIGITAL COMPOSITE IMAGES, INCLUDING IMAGES OF SEVERAL OBJECTS | 2011 |
|
RU2458396C1 |
| DETECTION OF IMAGES OF SCREEN ON IMAGES OF DOCUMENTS | 2014 |
|
RU2595557C2 |
| IMAGE PROCESSING SYSTEM AND METHOD USING DEEP NEURAL NETWORKS | 2018 |
|
RU2743931C1 |
| CLASSIFICATION OF DOCUMENT IMAGES BASED ON PARAMETERS OF COLOUR LAYERS | 2015 |
|
RU2603495C1 |
| RECONSTRUCTION OF THE DOCUMENT FROM DOCUMENT IMAGE SERIES | 2017 |
|
RU2659745C1 |
