Colour image segmentation based on fuzzy rules derived from entropy measurement

• Autorzy: Tarnawski W. , Merta A.
• Konferencja: Conference on Computer Graphics and Image Processing, GKPO' 2000 (6 ; 15-19.05.2000 ; Polesie, Poland)
• Języki publikacji: EN

Abstrakty

EN

This paper introduces a new colour image segmentation method based on fuzzy rules derived from measures of fuziness (fuzzy entropy) of an input image. The proposed supervised process of segmentation proceeds in two steps: first - user selects the region including pixels " more-or-less" connected with image objects, at the second step - based on generated and tuned fuzzy membershipes functions - the system automatically generate fuzzy rules and on the basic of them finds in the whole image all pixels which are similar to pixels from selected region. By combining the statistical approach connected with entropy measurement with uncertian or ambigous data selected by the user we received very high repaetability of results indepedently from the degree of precision in defining the region of features. Results are presented for using the mentioned method in analysing of medical images.

Słowa kluczowe

EN

colour image segmentation; fuzzy sets and rules; maximum entropy principle

• Wydawca: Institute of Information Technology of the Warsaw University of Life Sciences – SGGW
• Czasopismo: Machine Graphics and Vision
• Rocznik: 2000
• Tom: Vol. 9, No. 1/2
• Strony: 487--495
• Opis fizyczny: Bibliogr. 5 poz.

Twórcy

autor

Tarnawski W.

autor

Merta A.

• Division of Systems and Computer Networks, Faculty of Electronics Wroclaw University of Technology, Poland,