Firefly Algorithm based An Enhanced Brain Tumor Segmentation and Classification System

The proper segmentation and classification of a patient’s brain tumour is an essential step in arriving at an appropriate diagnosis and developing a treatment strategy for that patient’s brain tumour. In this research, we present a unique method for the segmentation and classification of brain tumours by combining the K-means clustering with the Firefly Algorithm (FA) for the tumour region segmentation and then after Maximally Stable Extremal Regions (MSER) feature based Convolutional Neural Network (CNN) is used to train the proposed an Enhanced Brain Tumour Segmentation and Classification (EBTSC) System. Both of these algorithms are used separately and the MRI (Magnetic Resonance Imaging) data from the brain are clustered using the K-means method, which separates the tumour from healthy brain tissue and organises the clusters into different areas. After then, the Firefly algorithm is used in order to perfect the segmentation procedure, which ultimately leads to improved precision and consistency in the findings. The proposed methodology consists of the following steps: First, the brain MRI data is pre-processed to enhance the image quality and remove noise. Next, the K-means clustering algorithm with FA is utilized to initially segment the brain MRI into several clusters. After obtaining the segmented regions, MSER feature extraction techniques are applied to extract relevant features from the segmented tumor region. These features are then fed into a classification model, such as a CNN, to classify the tumor into different types (e.g., benign or malignant). The classification model is trained on a labeled dataset to learn the patterns and characteristics of different tumor types, enabling accurate classification of unseen tumour cases and the results were compared with existing state-of-the-art methods. The experimental results demonstrated that the combination of K-means clustering and the Firefly algorithm achieved superior segmentation accuracy and classification performance, outperforming other existing techniques.