Towards a Method of Correcting Bias Introduced by the Use of the Error Fitting Method in Conjunction with a Maximum Entropy Processing Algorithm

Abstract

High resolution dark field electron micrographs are corrupted by large amounts of noise arising mainly from the scattering of electrons by the thin carbon film that is used to support the samples that are to be imaged. Processing to produce an image from which a structure can be deduced has in the past relied on methods such as filtering with crystalline arrays, contrast enhancement and the averaging of a number of images of similar structures. These methods however are rather arbitrary in their application and often lead to a reduction in the spatial resolution of resultant images. A maximum entropy processing method applied to digitised micrographs has been investigated to see whether significant image enhancement could be achieved without either subjective decisions as to the degree of image processing or a reduction in the spatial resolution in the final image.