Improved Support Vector Machine Kernel Utilising Hybrid Evolutionary Techniques for Medical Image Classification

Abstract

Digital imaging advancements have led to the creation and storage of massive medical picture archives. Images captured by X-ray, CT, MRI, PET, and ultrasound machines provide crucial anatomical and functional data for use in diagnosis, research, and education. Digital medical photos provide a large and diverse medical database. Good image retrieval systems are essential for the effective search of large data sets. Image Searching Based on Content CBIR is currently often used in the medical imaging retrieval industry to find pictures in the database that are comparable to the query image. Feature extraction and matching are the two main components of CBIR. Extracting characteristics from a picture in order to create a unique representation is known as "feature extraction." Extend Feature vectors are a kind of visual data that includes information like colour, texture, and shape. Both query pictures and stored images in databases go through this procedure. In order to find a suitable match, Feature Matching compares certain characteristics of each picture in the database. The most challenging aspect of CBIR systems is the incorporation of flexible methods for processing pictures with a wide range of attributes and classifications. This study investigates the challenge of extracting medical photos from a large and diverse collection. The primary goal of this study is to compare and contrast existing content-based image retrieval classification techniques, and to propose a new classification methodology.