FIELD: medicine.
SUBSTANCE: invention relates to medical field and can be used for detection and classification of colour images of nucleated blood cells and bone marrow when forming a diagnostic conclusion in oncohematology. Disclosed is a method for detecting and classifying colour images of nucleated blood cells and bone marrow when forming a diagnostic conclusion in oncohematology, characterized by the following stages: obtaining images using an automatic analyser of biomedical images of blood and bone marrow preparations; annotation of the obtained images by an expert with the establishment of marks on them indicating the type of the nucleated cell and forming a large and varied annotated data set containing images of different types of conditions of nucleated cells of bone marrow; extraction of square images (tiles) of input size of neural network from source images; creation of a neural network with architecture, for classification of images of blood cells and bone marrow; adjustment of hyperparameters of a neural network; calculation of the number of found nucleus-containing elements with their subsequent assignment based on the data of the trained neural network to the corresponding hematopoiesis germ and forming a diagnostic conclusion; visual inspection of an expert with subsequent correction of the result.
EFFECT: invention provides higher accuracy of nucleated blood and bone marrow cells when forming a diagnostic conclusion in oncohematology for automatic analysers of biomedical images.
1 cl, 7 dwg
Title | Year | Author | Number |
---|---|---|---|
METHOD OF DIAGNOSING ONCOLOGICAL DISEASES OF THE BLOOD | 2022 |
|
RU2803281C1 |
METHOD FOR SELECTING IMAGES AT HIGH MAGNIFICATION USING NEURAL NETWORKS IN THE STUDY OF A BONE MARROW SMEAR AND A SYSTEM FOR IMPLEMENTING THE METHOD | 2022 |
|
RU2793060C1 |
METHOD FOR ISOLATING AND CLASSIFYING BLOOD CELL TYPES USING DEEP CONVOLUTION NEURAL NETWORKS | 2019 |
|
RU2732895C1 |
FUEL COMBUSTION MODES MONITORING SYSTEM BY MEANS OF TORCH IMAGES ANALYSIS USING CLASSIFIER BASED ON CONVOLUTIONAL NEURAL NETWORK | 2018 |
|
RU2713850C1 |
DIAGNOSTIC TECHNIQUE FOR ACUTE CORONARY SYNDROME | 2020 |
|
RU2733077C1 |
METHOD FOR FORMING A TRAINING SAMPLE FOR CONTROL SYSTEMS OF UNMANNED ELECTRIC TRAINS | 2022 |
|
RU2785704C1 |
HARDWARE-SOFTWARE COMPLEX DESIGNED FOR TRAINING AND (OR) RE-TRAINING OF PROCESSING ALGORITHMS FOR AERIAL PHOTOGRAPHS OF THE TERRITORY FOR DETECTION, LOCALIZATION AND CLASSIFICATION UP TO TYPE OF AVIATION AND GROUND EQUIPMENT | 2020 |
|
RU2747044C1 |
METHOD FOR PROVIDING COMPUTER VISION | 2022 |
|
RU2791587C1 |
METHOD OF RESTORING WHITE BALANCE ON AN IMAGE | 2023 |
|
RU2837078C2 |
METHOD FOR RECOGNITION OF THE STRUCTURE OF BLOOD BLAST NUCLEI AND BONE MARROW USING LIGHT MICROSCOPY IN COMBINATION WITH COMPUTER DATA PROCESSING FOR THE DIAGNOSIS OF B- AND T-LINEAR ACUTE LYMPHOBLASTIC LEUKEMIA | 2017 |
|
RU2659217C1 |
Authors
Dates
2025-03-28—Published
2024-03-04—Filed