“Analyzing the Performance of SVM-ACO Classifier and Hybrid Optimization Techniques in MRI Brain Tumor Segmentation for Early Prognosis”
Year: 2023
Type: article
Abstract: In MRI images, automated methods are used to generate images of the human body's interior organs for clinical study. A key component of many medical evaluations is image segmentation. There are manual, automatic, and semi-automated methods for segmenting a region. The technique presents a hybrid process for classifying and identifying brain tumours using MRIs (magnetic resonance imaging). The first step of the suggested method corresponds with the pre-processing of MR images, which includes the denoise filter, cranium stripping, and other techniques. The following stage is FE (feature extraction), which is different from MR images of brain tumours, of brain MRI using Local Binary Pattern along with a histogram. Brain tumours are categorised using the SVM (Support Vector Machine) based on many types of characteristics. The third phase focuses on a decision compound that classifies Support Vector Machine and (ACO) Ant colony optimisation is dependent on decision rates concerned with confidence criteria to support the final distinctive result. Since a genuine human brain and artificial MRI data set were used in the trials, 158 MR images total normal MRIs and 100 aberrant MRIs were obtained. On both training and testing data images, it was determined that the categorization accuracy was 98.99% accurate.. The classifier differentiates the cancer with rather good accuracy and provides the radiologist with confirmation. The early detection of brain tumours and other brain disorders, as well as the planning of their treatment, depend on the medical image segmentation of brain images. MRI segmentation is a labor-intensive task best left to medical professionals. For the purpose of early diagnosis and treatment of brain tumours, this research needs to provide some fresh hybrid picture segmentation and classification techniques. In order to increase the reliability of the fuzzy c-means clustering (FCM) perspective for segmenting the images, local spatial information is typically supplied to an objective function.
Authors: Warjurkar, Sarvesh, Ridhorkar, Sonali
Institution: Sant Gadge Baba Amravati University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7217433734