Levy Flight and Chaos Theory-Based Gravitational Search Algorithm for Image Segmentation

Image segmentation is one of the pivotal steps in image processing due to its enormous application potential medical analysis, data mining, and pattern recognition. In fact, process splitting an into multiple parts order provide detailed information on different aspects image. Traditional techniques suffer from local minima premature convergence issues when exploring complex search spaces. Additionally, these also take considerable runtime find optimal pixels as threshold levels are increased. Therefore, overcome computational overhead problems multilevel thresholding process, a robust optimizer, namely Levy flight Chaos theory-based Gravitational Search Algorithm (LCGSA), employed perform COVID-19 chest CT scan images. LCGSA, exploration carried out by flight, while chaotic maps guarantee exploitation space. Meanwhile, Kapur’s entropy method utilized for segmenting various regions based pixel intensity values. To investigate performance ten versions firstly, several benchmark images USC-SIPI database considered numerical analysis. Secondly, applicability LCGSA solving real-world examined using imaging datasets Kaggle database. Further, ablation study considering ground truth Moreover, qualitative quantitative metrics used evaluation. The overall analysis experimental results indicated efficient over other peer algorithms terms taking less time providing values quality metrics.