Scholarly record

CLASSIFICATION OF MULTICHANNEL IMAGES BASED ON CELLULAR PROCESSES

Р. А. Томакова

Abstract

The aim of the work is the analysis of multichannel images intended for medical research related to the classification of radiographs. Two methods of descriptor formation are proposed as a basis for constructing the rules of classification of multichannel raster images. Based on these descriptors, two groups of classifiers are built with the subsequent aggregation of decisions made by these classifiers. Descriptors are generated based on the analysis of the contours of segment edges of the corresponding raster in those channels in which the image has good spatial resolution. To classify the selected contours, rastersare used in channels with good resolution in the region of spatial frequencies or in the region of the electromagnetic spectrum. The use of different-scale windows in each channel allows you to form plenty of classifiers for one channel with the subsequent aggregation of solutions both within the channel and between the channels. As a result, a network structure of classifiers (cellular classifiers) is formed, its parameters are determined through training, based on expert assessments or hybrid methods. The result of the research is the development of efficient algorithms for processing and analyzing multi-channel images. The structure of models based on cellular processors using neural networks has been established. The developed structures can be adapted to the specific features of the image and allow you to implement the classification of objects in medical images in real time. We have come to the conclusion about the possibility of applying the method to building an intelligent decision-making system for all types of processed multi-channel raster images.

Publication Impact Profile

• Captures
• Mendeley - Readers: 4

PLUMX see details

DOI resolver: doi.org/10.5593/sgem2019/2.1/s07.019

Publication details

Back to publication list