Title :
Multi-feature edge extraction with local fractal features
Author :
Zhuang, Xiaodong ; Meng, Qingchun
Author_Institution :
Comput. Sci. Dept., Ocean Univ. of China, Qingdao, China
Abstract :
The local fractal dimension (LFD) is proposed to extract local fractal features of gray-scale images. The definition of LFD is an extension of the pixel-covering method, based on the wavelet-based fractal dimension. Multi-feature edge extraction is implemented with the local fractal features and the image intensity discontinuities. The experimental results show that the LFD is an important feature of edge areas and the performance of the multi-feature edge extraction method is better than traditional gradient-based edge extraction methods.
Keywords :
edge detection; feature extraction; fractals; edge areas LFD; feature extraction; gray-scale images; image intensity discontinuities; local fractal dimension; local fractal features; multifeature edge extraction method; pixel-covering method; wavelet-based fractal dimension; Computer science; Data mining; Fractals; Geometry; Gray-scale; Image edge detection; Image texture analysis; Linear regression; Pixel; Wavelet transforms;
Conference_Titel :
TENCON 2003. Conference on Convergent Technologies for the Asia-Pacific Region
Print_ISBN :
0-7803-8162-9
DOI :
10.1109/TENCON.2003.1273324
