论文部分内容阅读
根据图像的统计信息,在Markov随机场(MRF)的理论基础上,构造了一种新的非线性算子即统计算子,提出了基于该算子的一种非线性提升小波分析方法.并根据MRF的条件概率分布,在理论上证明了采用基于统计算子的非线性提升小波变换可使图像变换后,在无量化失真的前提下,提高高频子带的零高频系数.将该方法与现有的几种非线性形态学小波分析方法以及S+P变换和JPEG2000采用的5/3和9/7线性提升小波变换进行了不同图像的测试分析,实验结果显示,利用这种基于统计算子的提升小波分析方法对医学图像和混合文档图像变换后可取得较低的加权熵.
According to the statistical information of the image, a new non-linear operator (statistical operator) is constructed based on the Markov random field (MRF) theory and a nonlinear lifting wavelet analysis method based on the operator is proposed. According to the conditional probability distribution of MRF, it is theoretically proved that using the non-linear lifting wavelet transform based on statistical operator can improve the zero-high frequency coefficients of high-frequency subband under the premise of no quantization distortion after image transformation. Methods Compared with the existing methods of nonlinear morphological wavelet analysis, S + P transform and 5/3 and 9/7 linear lifting wavelet transform used by JPEG2000, the experimental results show that this method based on Statistical operator lifting wavelet analysis can obtain lower weighted entropy for medical images and mixed document images.