A novel bio-inspired optimization algorithm for medical image restoration using Enhanced Regularized Inverse Filtering
Yasser Radouane Haddadi, Boualem Mansouri, Fatima Zohra Driss Khodja · Research on Biomedical Engineering · 2023
In medical imaging acquisitions, the image degradation process is composed of the blur and noise resulting in the imaging framework. Regularization methods are essential to obtain meaningful solutions represented as images restored after degradation. The regularization estimation in this approach controls the performance and effectiveness of image restoration operations. In this paper, we address the problematic of improving medical image restoration with regularization estimation. The application of a novel approach to medical image restoration is named ERIFGQPSO, which is based on the Gaussian Quantum-Behaved Particle Swarm Optimization algorithm and Enhanced Regularized Inverse Filter to control the medical image deblurring and denoising restoration operation. Our proposed ERIFGQPSO method provides the most accurate regularization parameter estimation. We achieved a Restoration Performance Ratio of 98.33, compared with other state-of-the-art restoration experimental methods. The application of our proposed approach to solve the regularization estimation issue is efficient for medical image restoration. It can be extended to several medical image-processing applications to improve image quality, sharpness, and operational effectiveness.