Wavelet transformation

This paper considers possible approaches to improving the quality of image signal formation both in the light and in the dark period of the day, as well as reducing the influence of noise, interference and artifacts on the characteristics of image signals. It is proposed to use the wavelet domain for the analysis of thermal and image signals with their subsequent possible complexation. The main features of the formation of such signals are indicated.

In the research a method for segmentation of partially-blurred images using the wavelet transformation, particularly coiflet of the order L=3 is presented. The entropy is used as a segmentation criterion based on wallet transformation. K-means method is used for image segmentation. Developed method was tested and has shown good results of his work; it correctly performs more than 82 % of pixels of image, and in many individual cases, more than 90 %.

In the paper, a classification method is proposed. It is based on Combined Swarm Negative Selection Algorithm, which was originally designed for binary classification problems. The accuracy of developed algorithm was tested in an experimental way with the use of microarray data sets. The experiments confirmed that direction of changes introduced in developed algorithm improves its accuracy in comparison to other classification algorithms.