Retrospective Shading Correction: Problem, Methods, and Evaluation

In this paper we address the problem of retrospective shading correction, which is a necessary pre-processing step in many tasks of image analysis. Eight algorithms are implemented and compared. The performance of the algorithms was analysed firstly; on three sets of differently structured synthetic shaded and shading-free images, and secondly; on two sets of real microscopical images acquired by two different acquisition set-ups. The results show that the entropy minimisation (EMI) method outperforms the other methods in terms of the reduction of true intensity variations and preservation of intensity characteristics of the shading-free images. The strength of EMI is especially apparent when applied to images containing large-scale objects.