3 Review(s)

Image segmentation is a vital yet challenging process that partitions images into meaningful regions to facilitate applications ranging from medical diagnostics and satellite remote sensing to robotics and agricultural monitoring. Still, it is often hampered by noise, low contrast, and intensity inhomogeneities introduced during acquisition. To address these issues, this book focuses on two principal techniques Multilevel Thresholding (MT) and Level Set Functions (LSF) and enhances them with modern optimization strategies: first, by replacing histogram‐only threshold selection with energy‐curve analysis combined with Otsu’s method and the Harmony Search Algorithm in the MTHSAE approach, which outperforms PSO and EMO in PSNR and fitness metrics on natural images; next, by applying EMO‐driven Otsu thresholding on energy curves for noise‐robust segmentation; then, by integrating MT with HSA plus morphological operations for precise white‐blood‐cell extraction in microscopic imagery, where the green channel yields the best MSE; after that, by pre‐processing low‐contrast medical and satellite images with HSA‐based bi‐histogram equalization before SBGFRLS level‐set segmentation in LSFBHEHS, achieving higher Dice indices; and finally, by introducing MTEMOE a context‐sensitive, multi‐thresholding scheme using both Kapur’s and Otsu’s criteria with EMO for color images, which delivers superior PSNR, SSIM, FSIM, PRI, VOI, and fitness scores compared to eight other histogram‐based algorithms.