FIELD: artificial intelligence; digital image processing.
SUBSTANCE: technical result is achieved due to: obtaining images using a photofixation device or image accumulation in the form of a database with an image class marking, transforming images using a spike neural network with synaptic plasticity into a universal new feature space, training and classification of a new feature space of images using a machine learning algorithm on a computing device and outputting the result on the display device. Invention provides separation of image classification process into stages, having the possibility of being implemented in the form of a microelectronic chip to form a new feature space and a computing device for image classification using different versions of the latter (microcomputer, personal computer, server).
EFFECT: automation of image classification due to formation of feature space using neuromorphic algorithm.
2 cl, 8 dwg
Title | Year | Author | Number |
---|---|---|---|
METHOD AND DEVICE FOR NEURAL TIME CODING, TRAINING AND RECOGNITION | 2012 |
|
RU2597504C2 |
NEURON MODEL BASED ON DENDRITIC COMPUTING | 2021 |
|
RU2777262C1 |
RESTORATION OF THE SENSORIMOTOR FUNCTION USING A NEUROPROSTHETIC SYSTEM AND METHOD FOR IMPLEMENTATION THEREOF | 2020 |
|
RU2772637C2 |
NEAR-REAL IMPULSE NEURON | 2015 |
|
RU2598298C2 |
METHOD FOR FORMING A SYNAPTIC MEMRISTOR BASED ON A NANOCOMPOSITE OF METAL-NONSTECHOMETRIC OXIDE | 2017 |
|
RU2666165C1 |
INFORMATION PROCESSING METHOD IN PULSED NEURAL NETWORK | 2019 |
|
RU2709168C1 |
SYSTEM FOR OPERATIONAL IDENTIFICATION OF MARINE TARGETS BY THEIR INFORMATION FIELDS BASED ON NEURO-FUZZY MODELS | 2021 |
|
RU2763384C1 |
METHOD FOR GENERATING MATHEMATICAL MODELS OF A PATIENT USING ARTIFICIAL INTELLIGENCE TECHNIQUES | 2017 |
|
RU2720363C2 |
SYSTEM FOR DETECTING AND CLASSIFYING NAVAL TARGETS BASED ON NEURAL NETWORK TECHNOLOGIES AND ARTIFICIAL INTELLIGENCE ELEMENTS | 2021 |
|
RU2780607C1 |
METHOD FOR TRAINING A CLASSIFIER DESIGNED FOR DETERMINING THE CATEGORY OF A DOCUMENT | 2017 |
|
RU2672395C1 |
Authors
Dates
2024-11-18—Published
2023-12-02—Filed