GLCM Features and Fuzzy C Means Clustering-Based Brain Tumor Detection in MR Images

Identifying and categorizing human brain tumors are labour-intensive activities, yet they are crucial for any doctor. There is a growing trend toward using computer-assisted diagnosis (CAD) to improve diagnostic abilities and raise detection accuracy to the highest possible levels. Variability in picture modality, contrast, tumor kind, and other characteristics makes brain tumor segmentation a difficult problem to solve despite years of study. Though there are many excellent works accessible, there is still a need for the development of efficient and precise ways for tumor segmentation with MR brain images. To address this gap in the literature, researchers have developed a new method for detecting human brain cancers that combines a template-based K-means (TK) algorithm, superpixels, and principal component analysis (PCA) to achieve faster detection rates while requiring less processing time overall. Super pixels and PCA can be utilized to identify the most useful information for the early detection of brain cancers. Then, a filter is used on the enhanced image to boost precision. Finally, the TK-means grouping technique is considered for subdividing the pictures for brain tumor identification. Super pixel-based feature extraction, on the other hand, leads to subpar segmentation results since it relies on region-based feature calculation rather than attempting to extract every possible feature from the brain pictures. Once the images have been improved by converting colour into grayscale, the Gray Level Co-Occurrence Matrix (GLCM) approach is utilized to extract the five statistical texture parameter features. Reduce the file size of an image by using a dimensionality reduction technique like the Independent Component Analysis (ICA) model. Finally, brain tumors are identified, and their segmentation is performed with the help of Fuzzy C Means clustering (FCM). The results obtained from a broad set of images further demonstrate the usefulness of the suggested model for recognizing the sizes and shapes of brain tumors.