Modified Remora Optimization Algorithm for Global Optimization and Multilevel Thresholding Image Segmentation
چکیده
Image segmentation is a key stage in image processing because it simplifies the representation of and facilitates subsequent analysis. The multi-level thresholding technique considered one most popular methods efficient straightforward. Many relative works use meta-heuristic algorithms (MAs) to determine threshold values, but they have issues such as poor convergence accuracy stagnation into local optimal solutions. Therefore, alleviate these shortcomings, this paper, we present modified remora optimization algorithm (MROA) for global tasks. We used Brownian motion promote exploration ability ROA provide greater opportunity find solution. Second, lens opposition-based learning introduced enhance search agents jump out To substantiate performance MROA, first 23 benchmark functions evaluate performance. compared with seven well-known regarding accuracy, speed, significant difference. Subsequently, tested quality MORA on eight grayscale images cross-entropy objective function. experimental metrics include peak signal-to-noise ratio (PSNR), structure similarity (SSIM), feature (FSIM). A series results proved that MROA has advantages among algorithms. Consequently, proposed promising method problems segmentation.