METHOD OF DETECTING NECESSITY OF STANDARD LEARNING FOR VERIFICATION OF RECOGNIZED TEXT Russian patent published in 2018 - IPC G06K9/66 G06F17/27 

Abstract RU 2641225 C2

FIELD: physics.

SUBSTANCE: method of the text analysis includes: analysis performed by the verification user of the recognized text obtained in the process of document image recognition where the verification involves changing the user-defined incorrect character by the user-defined correct symbol; identifying the similar changes of the first incorrect symbol on the first correct symbol; and initiating the process of learning the recognition standard based on the identified similar changes, where the recognition standard is learned to recognize a particular symbol and is used when recognizing characters in the document image to produce a recognized text.

EFFECT: increasing the overall accuracy of the document recognition.

18 cl, 6 dwg

Similar patents RU2641225C2

Title Year Author Number
HANDWRITING RECOGNITION USING NEURAL NETWORKS 2020
  • Andrey Upshinskiy
RU2757713C1
CHARACTER RECOGNITION USING A HIERARCHICAL CLASSIFICATION 2018
  • Aleksey Alekseevich Zhuravlev
RU2693916C1
VERIFICATION OF OPTICAL CHARACTER RECOGNITION RESULTS 2016
  • Kalyuzhny Aleksey Ivanovich
RU2634194C1
NEURAL NETWORK TRAINING BY MEANS OF SPECIALIZED LOSS FUNCTIONS 2018
  • Aleksey Alekseevich Zhuravlev
RU2707147C1
CONTENT-BASED DOCUMENT IMAGE CLASSIFICATION 2014
  • Smirnov Anatoly Anatolyevich
  • Panferov Vasily Vladimirovich
  • Isaev Andrey Anatolyevich
RU2571545C1
OPTICAL CHARACTER RECOGNITION SYSTEM AND METHOD, REDUCING PROCESSING TIME FOR IMAGES POTENTIALLY NOT CONTAINING CHARACTERS 2014
  • Chulinin Yuri Georgievich
RU2571616C1
INCOMPLETE STANDARDS 2013
  • Khintsitskij Ivan Petrovich
  • Masalovich Anton Andreevich
  • Isaev Andrej Anatolevich
RU2641452C2
OPTICAL CHARACTER RECOGNITION USING SPECIALIZED CONFIDENCE FUNCTIONS, IMPLEMENTED ON THE BASIS OF NEURAL NETWORKS 2018
  • Aleksey Alekseevich Zhuravlev
RU2703270C1
METHODS AND SYSTEMS OF EFFECTIVE AUTOMATIC RECOGNITION OF SYMBOLS USING A MULTIPLE CLUSTERS OF SYMBOL STANDARDS 2014
  • Chulinin Yurij Georgievich
RU2648638C2
DIFFERENTIAL CLASSIFICATION WITH MULTIPLE NEURAL NETWORKS 2017
  • Zhuravlev Aleksey Alekseevich
  • Vladimir Rybkin
  • Anisimovich Konstantin Vladimirovich
  • Davletshin Azat Aydarovich
RU2652461C1

RU 2 641 225 C2

Authors

Krivosheev Mikhail Viktorovich

Kolodkina Natalya Aleksandrovna

Makushev Aleksandr Sergeevich

Dates

2018-01-16Published

2014-01-21Filed