An improved contrast equalization technique for contrast enhancement in scanning electron microscopy images

摘要

In scanning electron microscopy (SEM), an intensive beam of electrons is utilized to generate high-quality images that can be used across scientific disciplines including biology and medicine. However, the images produced by different SEM devices are often degraded by a low-contrast effect as a result of various existing limitations. Hence, an improved contrast equalization technique is proposed in this article to ameliorate the contrast of SEM images and provide better visual quality results. The proposed technique consists of two major phases; the first includes a two-step regularization process used to regularize the intensities of the input image, while the second includes a two-step mapping process used to further improve the contrast and to remap image intensities to their natural dynamic range. The proposed technique is tested with many real-degraded and synthetic-degraded SEM images, and its results are compared with six well-known contrast enhancement techniques. Accordingly, the comparison results are evaluated using four advanced image quality assessment metrics. Intensive experiments and comparisons exhibited the favorability of the proposed technique, in that it provided visually pleasing results with no visible flaws and outperformed the comparison techniques in terms of perceived quality and scored accuracy.