Image Edge Detection with Fuzzy Classifier.

作者: Carl G. Looney , Lily R. Liang , Ernesto G. Basallo

DOI:

关键词:

摘要: Our special fuzzy classifier operates on the set of eight features extracted from 3x3 neighborhood each pixel. These are magnitudes differences between that pixel and neighboring pixels. They input into inputs connect to two membership functions represent “white background” or “black edge.” The paradigm is simple, computationally efficient, has low sensitivity noise isotropic. Each in image mapped white black. yields bold black lines a background. benefits employing for edge detection its small computation, noise, isotropy easy modeling, We discuss these more detail after introducing methodology. Methodology. For center p5, graylevel difference p5 neighbors designated by X1, X2,...,X8 calculated X1=p1-p5 X2=p2-p5 X3=p3-p5 X4=p4-p5

参考文章(5)
F. Russo, A user-friendly research tool for image processing with fuzzy rules ieee international conference on fuzzy systems. pp. 561- 568 ,(1992) , 10.1109/FUZZY.1992.258702
F. Russo, G. Ramponi, Fuzzy operator for sharpening of noisy images Electronics Letters. ,vol. 28, pp. 1715- 1717 ,(1992) , 10.1049/EL:19921090
Carl G. Looney, Nonlinear Rule-based Convolution for Refocusing Real-time Imaging. ,vol. 6, pp. 29- 37 ,(2000) , 10.1006/RTIM.1998.0154
Carl G. Looney, Fuzzy and rule-based image convolution soft computing. ,vol. 51, pp. 209- 219 ,(2000) , 10.1016/S0378-4754(99)00118-4