摘要
针对分数低阶α稳定分布信号具有显著脉冲这一特性,参照最优小波包基选择的熵准则,提出了一种针对该类噪声的最优小波基选择新准则,解决了以往最优小波基选择的熵准则中,由于需计算二阶统计量而不适合分析分数低阶α稳定分布信号的问题。还得出了针对该类信号,在运用该新准则进行判别最优小波基时,只需对各个小波基函数在第1层尺度下的变换系数进行对比判别的结论,极大地减化了以往多尺度的判别过程。将新准则应用于医学超声图像消噪时,获得了较好的效果。
In light of the entropy criteria on selecting the optimal wavelet packet basis and regarding the evident pulse property of the factional lower order a-stable distribution signals, this paper puts forward a new kind of criteria for selecting the best wavelet basis in coping with this kind of noise and offers a solution to the inexact analysis of the previous ones as they require the calculation of the second-order statistics. It also concludes that wile using the new guideline to distinguish the optimal wavelet basis, it is necessary only to make comparisons to the alternating coefficients of the first-section in every wavelet basis functions. And when applying the new guideline is applied in medical ultrasound image denoising, the result is satisfying.
出处
《通信技术》
2008年第11期185-187,共3页
Communications Technology
基金
国家自然科学基金资助项目(60772037)
九江学院科研课题(08KJ46)。
关键词
超声图像
熵准则
分数低阶
稳定分布
最优小波基
ultrasound image
entroy
fractional lower order
stable distribution
optimal wavelet basis
