Optimizing lung cancer diagnosis using improved fungal growth optimizer-based medical image segmentation.
Authors
Affiliations (2)
Affiliations (2)
- Department of Information Technology, Faculty of Computers and Informatics, Zagazig University, Zagazig, 44519, Egypt. [email protected].
- Department of Information Technology, Faculty of Computers and Informatics, Zagazig University, Zagazig, 44519, Egypt.
Abstract
Medical image segmentation is one of the most important processes in computer-aided diagnosis. It plays a critical role in detecting and analyzing diseases since it isolates the region of interest in medical scans. Out of many techniques, multilevel thresholding is capable of segmenting complex images. It isolates important structures with multiple intensity thresholds. The effectiveness of multilevel thresholding depends on the optimization algorithm's capacity to identify the appropriate threshold levels. This paper proposes an effective Fungal Growth Optimizer (FGO) with Orthogonal Learning Strategy (OLS). We emphasize enhancing solution diversity and speeding up convergence in multilevel thresholding-based medical image segmentation. The proposed algorithm, referred to as OLFGO, is tested on a set of 20 chest CT images from an openly available lung cancer database. We use various threshold values (2, 4, 6, 8, 10, and 12) to test its performance at different levels of segmentation. To assess the segmentation performance, we use several evaluation criteria such as Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), Feature Similarity Index Measure (FSIM), Universal Quality Index (UQI) and Dice coefficient (DICE). Moreover, the results are compared with deep learning algorithm's outcomes. We have demonstrated the reliability and efficacy of the OLFGO algorithm after implementing OLS using the Congress on Evolutionary Computation (CEC 2022) benchmarks. The improvement significantly enhanced the performance of the algorithm. The experiments confirm the efficiency of orthogonal learning in enforcing FGO's ability towards effective and reliable medical image segmentation. We also employed the Friedman rank sum test to rank the performance of OLFGO against existing techniques. OLFGO was ranked first, which reconfirms its superior ability in image segmentation. The MATLAB implementation of the proposed OLFGO is available at : https://github.com/shimaa21magdy-sketch/Orthogonal-Learning-Fungal-Growth-Optimizer-for-Lung-CT-Segmentation .