FIELD: computing equipment for image processing.
SUBSTANCE: computerized method of training, involving the following operations: obtaining access to image data; machine learning algorithm is used, to obtain machine annotations of objects in one or more scales of said image; generating a viewing window configurable based on the magnification ratio; simulating manual annotation of an object at different scales and/or in different parts of each image scale; using simulated manual annotations as different training inputs into machine learning algorithm; quantifying changes in various resultant machine annotations of objects; increasing coefficient and spatial offset parameter based on identified simulated manual annotation and after receiving the input data, manual annotation of the object by the user is used as a training input to the machine learning algorithm.
EFFECT: technical result consists in improvement of annotation of objects in image due to machine learning algorithm.
20 cl, 7 dwg
| Title | Year | Author | Number |
|---|---|---|---|
| METHOD FOR INTERACTIVE SEGMENTATION OF OBJECT ON IMAGE AND ELECTRONIC COMPUTING DEVICE FOR REALIZING SAID OBJECT | 2020 |
|
RU2742701C1 |
| TRAINING OF NEURAL NETWORK MODEL | 2018 |
|
RU2788482C2 |
| SEGMENTATION AND PREDICTION OF TIME PATTERNS OF LOW LEVEL TRAILS | 2020 |
|
RU2783420C1 |
| TRAINING GAN (GENERATIVE ADVERSARIAL NETWORKS) TO CREATE PIXEL-BY-PIXEL ANNOTATION | 2019 |
|
RU2735148C1 |
| AUTOMATIC DETECTION AND EXTRACTION OF PREVIOUS ANNOTATIONS, RELEVANT FOR IMAGING STUDY, FOR EFFICIENT VIEWING AND REPORTING | 2013 |
|
RU2640009C2 |
| SEGMENTATION OF ANATOMICAL STRUCTURE BASED ON MODEL | 2016 |
|
RU2721078C2 |
| SYSTEMS AND METHODS OF SEGMENTING IMAGE USING CONVOLUTION NEURAL NETWORK | 2017 |
|
RU2720070C1 |
| INTEGRATED DESKTOP SOFTWARE FOR MANAGEMENT OF VIRUS DATA | 2010 |
|
RU2520423C2 |
| METHOD AND SYSTEM FOR SEARCHING GRAPHIC IMAGES | 2022 |
|
RU2807639C1 |
| IMAGE SEGMENTATION METHOD USING NEURAL NETWORK | 2017 |
|
RU2720440C1 |
