Multiverse optimization with information sharing and spiraling communication: Enhancing multithreshold image segmentation for colon cancer

Abstract The early diagnosis of colon cancer is crucial for enhancing patient survival rates. Computer-aided image segmentation technology offers a significant tool for assisting physicians in quickly and accurately identifying cancerous tissues, serving as an effective means of auxiliary diagnosis. However, traditional image segmentation algorithms often encounter limitations in their optimization capabilities when processing complex images, struggling to meet the requirements for efficiency and precision. This study introduces an improved multiverse optimization algorithm, which, by incorporating information-sharing search and spiraling communication and collaboration strategies, aims to enhance the search efficiency and solution quality of the algorithm, thereby more effectively addressing the challenges of colon cancer image segmentation. Initially, using the CEC 2017 benchmark function test set, the proposed algorithm was compared to 11 meta-heuristics and 11 state-of-the-art algorithms, respectively, achieving the best rankings of 2.22 and 2.72 for the Friedman test. Subsequently, the proposed algorithm was applied to the task of multithreshold segmentation of colon cancer images. Compared to existing algorithms, the proposed algorithm achieved higher quality image segmentation results and stronger robustness at both high and low thresholds using feature similarity index, peak signal-to-noise ratio, and structural similarity index. This validates the effectiveness and superiority of the proposed algorithm in addressing the challenges of colon cancer image segmentation.