Quantum computing-based metaheuristics for medical image segmentation
Ahmad Sajad Rather, Sujit Ranjan Das · 2025
In the realm of image processing, image segmentation serves as a vital stage, furnishing in-depth insights into diverse facets of an image by partitioning it into multiple zones based on pixel intensity. This division process enables a granular examination of different components within the image, thereby facilitating subsequent analysis and interpretation. However, conventional segmentation methods often face difficulties like local minimums and early convergence when navigating complex pixel search spaces. Additionally, these methods can be computationally intensive, especially as threshold levels increase. To overcome these challenges, we employed a robust optimization method known as Gaussian Quantum-behaved Particle Swarm Optimization (GQPSO) for multi-level thresholding. Inspired by classical Particle Swarm Optimization (PSO) methods and principles of quantum mechanics, GQPSO integrates a mutation operator that utilizes a Gaussian probability distribution. Unlike traditional PSO approaches, this allows more focused exploration of the solution space, enhancing the algorithm’s ability to identify optimal solutions and mitigate premature convergence to local optima. In evaluating the segmentation performance and versatility of GQPSO, we analyzed various COVID-19 chest CT scan datasets sourced from the Kaggle database. Our assessment utilized multiple image evaluation metrics, including Peak Signal to Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM), and Feature Similarity Index Measure (FSIM), to gauge the quality, symmetry, and consistency of the segmented output. Qualitative analysis was conducted using convergence curves, segmented graphs, colormap images, histogram maps, and box plots. Furthermore, we conducted a comparative analysis with several state-of-the-art heuristic algorithms. The comprehensive examination of experimental results revealed the superior performance of GQPSO in terms of computational efficiency and its ability to yield optimal values for image quality metrics. Moreover, the simulation results underscored the potential of GQPSO in efficiently evaluating the severity of COVID-19 disease, highlighting its effectiveness in solving various optimization problems.