| Abstract [eng] |
This work is dedicated for Krawchouk and Chebyshev polynomials. I tried to compare two types of polynomials in image analysis. These polynomials belong to discrete orthogonal polynomial family. Work is divided in two main parts: theoretical part and practical part. In theoretical part I introduced both polynomials, their moments and invariants. Also, I talked about image reconstruction and classification. In practical section I showed how polynomials deal with image reconstruction, classification and found very important feature of polynomials – image transformation using polynomials work as noise reduction filter. This is absolutely way of polynomials usage. This can be useful not only with images, but also with density functions and number matrices. Krawchouk polynomials showed better results in all these practical examples. So I am doing assumption that discrete Krawchouk polynomials is better in image analysis comparing to Chebyshev polynomials. |
