Title
Image Enhancement and Features Extraction of Electron Microscopic Images Using Sigmoid Function and 2D-DCT
Abstract
An innovative image enhancement and feature extraction technique that is based on the modified sigmoid function and two-dimensional DCT has been developed. The proposed technique uses a modified sigmoid function that accommodates the original microscopic input image characteristics. A novel block-based input value coupled with the modified sigmoid function is used in this proposed technique to provide good contrast enhancement of an image, resulting in localised contrast enhancement. Singular value decomposition played an important role after DCT because the singular value matrix determines the intensity values of the provided microscopic image. Changes in the singular values have an immediate impact on the intensity of the microscopic input image. The proposed methodology essentially converts the input picture into the SVD-DCT domain, normalises the singular value matrix, and finally reconstructs the enhanced image using inverse DCT. Simulation findings demonstrate that the proposed technique produces significantly superior improved results compared to other current approaches. Various essential characteristics of actinomycetes become evident once electron microscopic images are enhanced, such as long filaments, coils or spirals, rod shapes, and spore patterns. The presented method works successfully and efficiently for various bright and dark microscopic images.
Year
DOI
Venue
2022
10.1109/ACCESS.2022.3192416
IEEE ACCESS
Keywords
DocType
Volume
Feature extraction, Discrete cosine transforms, Image enhancement, Electron microscopy, Histograms, Matrix decomposition, Transforms, Enhancement, modified sigmoid function, SVD, adaptive histogram equalization, DCT
Journal
10
ISSN
Citations 
PageRank 
2169-3536
0
0.34
References 
Authors
0
5