FIELD: physics, computer engineering.
SUBSTANCE: invention relates to decoding images obtained from image sensors. Disclosed is a method of decoding images, which includes forming images based on information obtained from image sensors. Method comprises segmenting images based on use of a multi-dimensional neutral Kohonen map on a pyramidal scheme, with segmentation at each level of the pyramid of different-scale images - representations of the initial image and further combination of segmentation results. Further, method comprises classification of patterns on the image using a tree classifier which uses as simple classifiers feedforward neural networks without feedback, trained using a genetic algorithm using Kitano graph generation grammar for encoding the network structure.
EFFECT: high accuracy of recognising objects on an image.
1 dwg
| Title | Year | Author | Number |
|---|---|---|---|
| NAVIGATION COMBINED OPTICAL SYSTEM | 2018 |
|
RU2694786C1 |
| DEVICE FOR MULTI-ANGLE MULTI-POSITION SYNTHESIS OF A COMPLEX IMAGE OF THE EARTH'S SURFACE | 2020 |
|
RU2756904C1 |
| HARDWARE-SOFTWARE COMPLEX DESIGNED FOR PROCESSING AERIAL PHOTOGRAPHS IN VISIBLE AND FAR INFRARED BAND FOR DETECTION, LOCALISATION AND CLASSIFICATION OF BUILDINGS OUTSIDE OF LOCALITIES | 2020 |
|
RU2752246C1 |
| DEVICE FOR FORMING A COMPLEX IMAGE BASED ON THE COMBINATION OF INDIVIDUAL HETEROGENEOUS IMAGES WITH A DIGITAL MAP OF THE AREA | 2022 |
|
RU2789857C1 |
| METHOD FOR DECIPHERING ELECTROMYOSIGNALS AND APPARATUS FOR IMPLEMENTATION THEREOF | 2020 |
|
RU2762775C1 |
| METHOD FOR DIAGNOSING INFORMATION-CONVERTING ELEMENTS OF AIRCRAFT ON-BOARD EQUIPMENT BASED ON MACHINE LEARNING | 2022 |
|
RU2802976C1 |
| HANDWRITING RECOGNITION USING NEURAL NETWORKS | 2020 |
|
RU2757713C1 |
| METHOD OF RECOGNIZING TRAFFIC LIGHT SIGNALS AT NIGHT | 2021 |
|
RU2770068C1 |
| VIEW CLASSIFICATION-BASED MODEL INITIALISATION | 2014 |
|
RU2669680C2 |
| METHOD OF AUTOMATIC SEGMENTATION OF FLUOROGRAMM OF BREAST OF PATIENTS WITH PNEUMONIA | 2016 |
|
RU2629629C1 |
