Research Blog
Medical image analysis using soft computing is a new field of medical imaging that provides a great scope for applying engineering analysis and modeling for medical diagnostic and research purposes where the available data set is vague or fuzzy. Today, medical image analysis, obtained from X-rays, magnetic resonance imaging (MRI), computed tomography (CT), pathological images, etc. are carried out using advanced digital imaging techniques. As medical images are not equally and well illuminated, for more accurate diagnosis, fuzzy/intuitionistic fuzzy/Type 2 fuzzy set/ Neutrosophic set theory is being explored.
A fuzzy set considers uncertainty in the form of membership or degree of belongingness of an element in a set. Intuitionistic fuzzy set considers that there may be some hesitation while assigning membership degree to an element in a set and so the non-membership degree is not the complement of the membership degree as in the fuzzy set. So, an intuitionistic fuzzy set considers two degrees of uncertainty – membership and non-membership degree. Type 2 fuzzy set considers that the membership degree assigned to an element in a set may also be fuzzy, so it is a Type 2 fuzzy set. Significant contributions have been made in the field of medical image processing using intuitionistic fuzzy set (IFS), a Type 2 fuzzy set for accurate diagnosis with increased accuracy.
Images have imprecise gray levels/vague boundaries/information loss while mapping. But medical images contain more uncertainties due to vagueness in homogeneity of image segments or vague contrast between regions; so IFS/Type 2 fuzzy set that considers either more or different forms of uncertainties as compared to fuzzy sets have been used. There is very little work on medical image processing using IFS/Type 2 fuzzy sets and mathematical models are given using these sets.
A new expression of an intuitionistic fuzzy generator for contrast enhancement of mammogram images is given. A new method using a Type 2 fuzzy set that uses Hamacher t conorm to enhance the medical images has been developed. These algorithms perform better on non-uniformly illuminated images which will facilitate the physicians to visualize image structures.
Sometimes edges are not properly visible; so the edges of image structures are enhanced. Direct applying edge detectors sometimes fails to detect boundaries. An intuitionistic fuzzy edge enhancement method that uses a rank-ordered filter to highlight the edges of images has been used. Edge enhancement is performed to enhance the edges of blood vessels using Type II fuzzy set theory and promising results are obtained. This will help physicians to visualize the boundaries at an initial stage.
Recently, a novel method for segmenting lesions in low-contrast mammogram images has been suggested where after image enhancement, a novel fuzzy hyperbolozation method has been given and then lesions are segmented using fuzzy clustering method on interval type 2 fuzzy image. [Journal of Digital Imaging, 2021, International Journal of Imaging Systems and Technology, 2020].
Currently, neutrosophic sets for extracting lesions in mammogram images almost accurately using the clustering approach are being explored.
Visit our website: https://www.ncuindia.edu/
Authored by
Tamalika Chaira
Associate Professor,
Department of Computer Science and Engineering,
The NorthCap University
Comments
Post a Comment