Image preprocessing algorithms of pigmented skin lesions and their influence on feature vector in classification using fractal parameters
Fragment książki (Materiały konferencyjne)
MNiSW
15
WOS
| Status: | |
| Autorzy: | Goleman Ryszard, Stryczewska Henryka, Mańko Monika, Giżewski Tomasz, Znajewska-Pander Aleksandra, Placek Waldemar, Owczarczyk-Samonek Agnieszka |
| Dyscypliny: | |
| Aby zobaczyć szczegóły należy się zalogować. | |
| Wersja dokumentu: | Drukowana | Elektroniczna |
| Arkusze wydawnicze: | 0,5 |
| Język: | angielski |
| Strony: | 1 - 4 |
| Web of Science® Times Cited: | 0 |
| Scopus® Cytowania: | 0 |
| Bazy: | Web of Science | Scopus | IEEE Xplore |
| Efekt badań statutowych | NIE |
| Materiał konferencyjny: | TAK |
| Nazwa konferencji: | 9th International Conference ELMECO-9 " Electromagnetic Devices and Processes in Environment Protection " with 12th Seminar AoS-12 "Applications of Superconductors" |
| Skrócona nazwa konferencji: | ELMECO & AoS 2017 |
| URL serii konferencji: | LINK |
| Termin konferencji: | 3 grudnia 2017 do 6 grudnia 2017 |
| Miasto konferencji: | Nałęczów |
| Państwo konferencji: | POLSKA |
| Publikacja OA: | NIE |
| Abstrakty: | angielski |
| The fractal analysis method in medical diagnosis is a promising tool in evaluating image parameters regardless of the scale used. From a discussion among dermatologists, it is clear that the problem of classification the pigmented skin lesions is large and the commercially available tools consider only changes in the shape, color, and symmetry in the image. The problems of selecting the binarization threshold, the use of graphical filters and the reduction of the number of free variables are presented. A computer program was created for automated image processing and fractal analysis, primarily based on the implementation of selected image binarization methods and selected image filters. Furthermore, the impact of the binarization threshold on fractal parameters was investigated. Statistical analysis has shown that the variables as fractal parameters can be linearly dependent on each other, and thus it is possible to reduce the input vector for classification algorithms. |
