Comparative evaluation of retrospective shading correction methods

Abstract

SummaryBecause of the inherent imperfections of the image formation process, microscopical images are often corrupted by spurious intensity variations. This phenomenon, known as shading or intensity inhomogeneity, may have an adverse affect on automatic image processing, such as segmentation and registration. Shading correction methods may be prospective or retrospective. The former require an acquisition protocol tuned to shading correction, whereas the latter can be applied to any image, because they only use the information already present in an image. Nine retrospective shading correction methods were implemented, evaluated and compared on three sets of differently structured synthetic shaded and shadingfree images and on three sets of real microscopical images acquired by different acquisition set-ups. The performance of a method was expressed quantitatively by the coefficient of joint variations between two different object classes. The results show that all methods, except the entropy minimization method, work well for certain images, but perform poorly for others. The entropy minimization method outperforms the other methods in terms of reduction of true intensity variations and preservation of intensity characteristics of shadingfree images. The strength of the entropy minimization method is especially apparent when applied to images containing large-scale objects.

Retrospective shading correction algorithm based on signal envelope estimation

“…The objects of interest in Fig. 1d are of similar size and nature as the image that most algorithms failed to correct in [8]. The central column presents the surfaces that corresponded to the shading of the images.…”

“…However, a limitation of these methods is that they assume that the background is either darker or brighter than the objects of interest, and that these are limited in size and smaller than the background variations. In [8] several methods were compared and all filtering methods failed to correct images with large objects.…”

CAIMAN: An online algorithm repository for Cancer Image Analysis

“…In light microscopy, the variation may originate from uneven sample thickness, out-of-focus objects (in thick slices), or departure from Köhler illumination. When the only data available is the image itself or when the shading is caused by the object, the only way to remove shading is with retrospective algorithms as opposed to a prospective algorithms that require a calibration protocol and extra images [9]. The retrospective algorithm based on an envelope estimation algorithm presented in [2] was implemented in CAIMAN.…”

Registration of adaptive optics corrected retinal nerve fiber layer (RNFL) images

“…This is due to inherent imperfections of the image formation process such as non-uniform illumination, uneven spatial sensitivity of the sensor or camera imperfections [16]. It affects automatic image processing, such as segmentation, registration and characterization of retinal features [17]. We have previously described a novel wavelet based method of pre-processing for the correction of uneven illumination in AO flood retinal images [18].…”