Fractional Brownian motion, continuous everywhere and differentiable nowhere, offers a convenient modeling for irregular nonstationary stochastic processes with long-term dependencies and power law behavior of spectrum over wide ranges of frequencies. It shows high correlation at coarse scale and varies slightly at fine scale, which is suitable for and successful in describing and modeling natural scenes. On the other hand, man-made objects can be constructively well described by using a set of regular simple shape primitives such as line, cylinder, etc. and are free of fractal. Based on the difference, a method to discriminate man-made objects from natural
scenes is provided. Experimental results verify its efficiency.
Mandelbrot B B, Van Vess J. Fractional Brownian motion, fractional noises and apllications. SIAM[2]Review, 1968, 10(4): 422-437.[3]Pentland A. Fractal-based description of natural scense. IEEE Trans. on PAMI, 1984, PAMI-6:[4]1-674.[5]Dubuisson M P, Dubes R C. Efficacy of fractal features in segmenting image of natural textures[J].Pattern Recognition Letters.1994, 15(4):419-431[6]Chaudhuri B B, et al. Improved fractal geometry based texture segmentation technique. IEE Proc-E, 1993, 140(5): 233-241.[7]Maragos P, Sun F K. Measuring the fractal dimension of signals: Morphological covers and iterative optimization. IEEE Tans. on SP, 1993, SP-41(1): 108-121.[8]谢文录,谢维信.离散时间序列中的分形分析和参效提取.信号处理(已录用).[9]Keller J M, et al. Texture description and segmenation through fractal geometry. Computer Vi-[10]sion,Graphics and Image Processing, 1989, (45): 150-166.[11]Sarkar N, Chaudhuri B B. An efficient approach to estimate fractal dimension of texture images. Pattern Recognization Letters, 1992, 25(9): 1035-1041.