首页|Hurst exponent estimation of fractional surfaces for mammogram images analysis
Hurst exponent estimation of fractional surfaces for mammogram images analysis
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NSTL
Elsevier
The work presents a methodology to precise simulation and parameter estimation of multidimensional fractional Brownian motion (fBm). The simulation approach uses circulant embedding algorithm and solution of Poisson equation, while generalizing it to multiple dimensions. For estimation, a method using Wishart distribution and maximum likelihood is presented and verified on simulated data. Unlike approximate methods for generating multidimensional fBm and its Hurst exponent estimation, this approach shows unbiased results for all processes with short memory and majority of cases with long memory. The methodology is applied to mammography screening images to find significant differences between benign and cancerous breast lumps. (C) 2021 Elsevier B.V. All rights reserved.
NSTL
Elsevier
