Employing Kapur’s Entropy to Identify Multilevel Threshold Segmentation in MRI Scans of Brain Tumors Using the Bioinspired Walrus Optimization Algorithm

Segmentation of brain tumors using magnetic resonance imaging (MRI) is important for computerized diagnosis, which is important for early diagnosis and better treatment, improving patient survival. Multilayer thresholding (MLT) is important for image segmentation, especially for complex non-linear images. Although MLT may seem simple, it often requires a high level of goal work. Therefore, researchers are working on bionic optimization algorithms (BIOA) as an alternative. This article explores how to use the newly developed Walrus Optimization Algorithm (WaOA) to optimize brain color MRI images at different levels. This algorithm is inspired by the natural behaviors of walruses, such as feeding, migrating, avoiding threats, and facing predators. In this work, Kapur's entropy approaches are used as the objective function. To perform quantitative analysis, the effectiveness of WaOA is evaluated using a set of six brain tumors MRI images and then contrasted with the outcomes of seven additional cutting-edge algorithms, across 2, 3, 4, and 5 color levels. The experimental findings revealed that the proposed algorithm consistently outperforms other algorithms across different output metrics such as Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and Mean Squared Error (MSE). For qualitative analysis, the non-parametric Friedman ranking test is employed to distinguish notable variances among the alternative methods. The results indicate that the suggested approach outperforms other known algorithms and shows superiority in terms of convergence, precision, and robustness.