Denoising 2D Images Using Perona-Malik Anisotropic Diffusion Model

This study implements the Perona-Malik anisotropic diffusion model for denoising 2D images. This method represents an effective approach for noise reduction while maintaining edge preservation in digital images. The Perona-Malik anisotropic diffusion model is a partial differential equation-based image processing technique known for its excellent denoising capabilities, making it well-suited for this research. In the implementation, diffusion is applied to pixel values through iterative updates using diffusion coefficients controlled by gradient magnitude, suppressing noise while ensuring edges are neither blurred nor compromised. Key functions typically involve calculating gradient magnitudes, applying diffusion coefficients (such as g(∇I) = 1/(1 + (∇I/K)²)), and performing iterative updates using finite difference schemes. Experimental results demonstrate that this method not only effectively removes noise but also exhibits outstanding performance in preserving crucial image edge information.