Fuzzy image processing : a review and comparison of methods

• Autorzy: Tzafestas, S. , Raptis, S.
• Języki publikacji: EN

Abstrakty

EN

This paper presents a comprehensive review of fuzzy image processing methods. Specifically the following image processing problems are considered: (i) Image comprehension. (ii) image segmentation, (iii) image classification, (iv) image analysis, (v) image filtering, (vi) image understanding. In practice, an image cannot always be interpreted always exactly and perfectly. This is due to the existence of noise or the way the image is obtained, or, finally, to incorrect understanding of the image information content. These difficulties can be faced successfully through fuzzy logic and fuzzy reasoning. The field of image processing via fuzzy logic was initialed after Zadeh's 1965 seminar paper and is still expanding with new important results and applications. This paper is devoted to the treatment of still images, but some of the methods can be extended to the case of moving (video) images. The methods considered are critically discussed. Finally, a comparison of the effectiveness of (i) c-means, (ii) classical c-means, and (iii) adaptive clustering algorithms is made.

Słowa kluczowe

PL

współczynnik rozmycia; kwantyzacja wektorowa

EN

classifccation map; discrepancy measure; discrete cosine transform; edge relaxation methods; features space; Fuzzy Associative Matrix (F.A.M.); fuzzy c-means algorithm; fuzzy edge detection; fuzzy edge enharccernent; fuzzy inferencing; fuzzy linear vector quantizer; fuzzy matrix; fuzzy region; high level image processing; image structures; local image regions features; local intensity gradient; local shape

• Czasopismo: Image Processing & Communications
• Rocznik: 1999
• Tom: Vol. 5, no 1
• Strony: 3--24
• Opis fizyczny: Bibliogr. 18 poz.,Rys., wykr.,

Twórcy

autor

Tzafestas, S.

autor

Raptis, S.