Denoised Images with Various Noise Intensities Optimized using Particle Swarm Algorithm

In this study, we investigate the application of Particle Swarm Optimization (PSO) algorithm to optimize structural elements for denoising images corrupted with varying noise intensities. Our implementation begins by analyzing noisy images with different standard deviations of Gaussian noise. We establish optimal PSO parameters through parameter tuning, including swarm size (typically 20-50 particles), inertia weight (decreasing from 0.9 to 0.4), and acceleration coefficients (c1=c2=2.0). The fitness function evaluates PSNR values achieved by morphological filtering using candidate structuring elements. The PSO algorithm iteratively updates particle positions representing structuring element configurations until convergence. Our morphological denoising implementation employs MATLAB's imopen() and imclose() functions with the optimized structuring elements. Comparative analysis with wavelet denoising (using wavedec2 and waverec2 functions) and median filtering (medfilt2) demonstrates our method's superior performance in maximizing PSNR. This research provides novel approaches and methodologies for the image denoising field.