An Efficient Adaptive Algorithm for Electron Microscopic Image Enhancement and Feature Extraction

In this article, a block-based adaptive contrast enhancement algorithm has been proposed, which uses a modified sigmoid function for the enhancement and features extraction of electron microscopic images. The algorithm is based on a modified sigmoid function that adapts according to the input microscopic image statistics. For feature extraction, the contrast of the image is very important and authentic property by which this article enhances the visual quality of the image. In this work, for better contrast enhancement of image, a block based on input value, combined with a modified sigmoid function that is used as contrast enhancer provides better EMF values for a smaller block size. It provides localized contrast enhancement effects adaptively which is not possible using other existing techniques. Simulation and experimental results demonstrate that the proposed technique gives better results compared to other existing techniques when applied to electron microscopic images. After the enhancement of microscopic images of actinomycetes, various important features are shown, like coil or spiral, long filament, spore and rod shape structures. The proposed algorithm works efficiently for different dark and bright microscopic images.

Keywords: Adaptive Histogram Equalization, EMF Values, Image Enhancement, Modified Sigmoid Function