FIELD: medical computer technology.
SUBSTANCE: method used for research and diagnostics, in particular using digital microscopy systems and artificial neural networks, for detecting and classifying formed elements in a bone marrow preparation. A method for selecting images for analysis obtained at high magnification when examining a bone marrow smear, which is performed using an automated digital microscopic system in two stages, in which at the first stage a digital image of the bone marrow preparation is obtained at low magnification, then a digital image of the bone marrow preparation of the brain into the first neural network to select areas of interest is entered, at the second stage, each image of the field of view obtained at high magnification is entered into the second neural network to assign such an image an image suitability factor for analysis, then the most suitable images are entered into the third neural network for rejection and removal of empty images and images containing artifacts, after which the remaining images are introduced into the fourth neural network to identify the desired shaped elements in the images, then the images with the desired shaped elements are transferred to the fifth neural network for segmentation shaped elements using a mask from the background, after which the segmented elements are transferred to the sixth neural network for classifying the segmented elements. Moreover the automated digital microscopic system includes a computer containing a machine-readable medium with instructions and deep convolutional neural networks recorded on it, as well as at least one processor that provides instructions and neural networks for implementing the described method.
EFFECT: reduction in the operating time of an automated microscopic system, a reduction in the amount of memory of the information storage system, and a reduction in labor costs during diagnostics.
2 cl, 5 dwg
Title | Year | Author | Number |
---|---|---|---|
METHOD FOR DIGITIZING A BIOLOGICAL PREPARATION OF BONE MARROW USING NEURAL NETWORKS | 2022 |
|
RU2793054C1 |
METHOD OF DIAGNOSING ONCOLOGICAL DISEASES OF THE BLOOD | 2022 |
|
RU2803281C1 |
METHOD FOR SCANNING BIOLOGICAL SUBSTANCES WITH THREE TYPES OF LOCALIZATION | 2023 |
|
RU2804992C1 |
METHOD FOR MICROSCOPIC EXAMINATION OF A SAMPLE WHICH CONTAINS MICRO-OBJECTS WITH HETEROGENEOUS ZONES | 2006 |
|
RU2308745C1 |
SEGMENTATION OF HUMAN TISSUES IN COMPUTER IMAGE | 2017 |
|
RU2654199C1 |
METHOD FOR ISOLATING AND CLASSIFYING BLOOD CELL TYPES USING DEEP CONVOLUTION NEURAL NETWORKS | 2019 |
|
RU2732895C1 |
METHOD FOR IDENTIFYING AND CALCULATING CELLS IN HUMAN AND ANIMAL BIOLOGICAL MEDIA AND DEVICE FOR ITS IMPLEMENTATION | 2004 |
|
RU2303812C2 |
METHOD OF DETECTING RETINAL PATHOLOGY | 2022 |
|
RU2802558C1 |
METHOD FOR AUTOMATED BLOOD CELLS ANALYSIS THROUGH LEUKOCYTES DESCRIPTION BASED ON OPTICAL FEATURES OF NUCLEI STRUCTURES | 2014 |
|
RU2612007C2 |
METHOD FOR DETECTING DENTS ON VEHICLE BODY ELEMENTS AND DEVICE FOR IMPLEMENTING THIS METHOD | 2021 |
|
RU2763307C2 |
Authors
Dates
2023-03-28—Published
2022-12-07—Filed