Штучний інтелект

Анотація: The work is devoted to the description of the grayscale images fuzzy clustering method, which performs dynamic transformation of initial data based on singular decomposition (with automatic determining the most important columns of left singular vectors matrix) on each step. This approach may lead to segmentation sensitivity enhancement. Proposed method was described on the example of neuro-fuzzy clustering algorithm sFCM. The results of experimental researches were obtained after processing of real grayscale medical image. Better identification of objects of interest and whole image structure was achieved.

Ключові слова:

Посилання:

1. Gonsales, R. (2006) Tsifrovaya obrabotkaizobrazheniy. – M.: Tehnosfera.
2. Ahmetshina, L.G. (2008) Segmentatsiyanizkokontrastnyih izobrazheniy algoritmomgibridnoy klasterizatsii SOM-FCM. Systemnitekhnolohii, 2 (55), 34-40.
3. Egorov, A.A. (2011) Povyishenie byistrodeystviyametoda gibridnoy nechetkoy klasterizatsii za schetdinamicheskogo szhatiya razmernosti kartyiKohonena. Shtuchnyi intelekt, 54 (4), 489-497.
4. Egorov, A. (2015) Optimizatsiya yarkostiizobrazheniy na osnove neyro-fazzi tehnologiy. −Lambert.
5. Ahmetshina, L.G. (2015) Vizualizatsiya rezultatovnechetkoy klasterizatsii izobrazheniy na osnovesingulyarnogo razlozheniya. Vestnik Hersonskogonatsionalnogo tehnicheskogo universiteta, 3 (54),198-202.
6. Forsayt, Dzh. (1968) Chislennoe reshenie sistemlineynyih algebraicheskih uravneniy.  M.: Mir.