Farrow N A, Ottensmeyer F P
Ontario Cancer Institute, University of Toronto, Ontario, Canada.
Scanning Microsc Suppl. 1988;2:75-81.
A maximum entropy algorithm is described which not only fits a model to the data consistent with the size of the noise and the maximum entropy principle but also distributes the residuals between the data and the model in a way consistent with the noise in the data having been generated by a random gaussian process. The results of applying the algorithm to profiles of electron micrographs, electron micrographs and model data is presented. The algorithm is found to achieve various degrees of signal to noise ratio enhancement. Preliminary results show that the spatial resolution is not suppressed. A biassing artifact is described.
描述了一种最大熵算法，该算法不仅能使模型与符合噪声大小和最大熵原理的数据相拟合，还能以一种与数据中的噪声由随机高斯过程产生相一致的方式，将残差分布在数据和模型之间。给出了将该算法应用于电子显微镜图像轮廓、电子显微镜图像和模型数据的结果。发现该算法能实现不同程度的信噪比增强。初步结果表明空间分辨率未受抑制。描述了一种偏差伪像。
