Image Denoising Based on Reaction-Diffusion Equation and Curvelet Transform
To reduce the pseudo-Gibbs effects and the “curvelet like” aliased curves resulted from using curvelet transform for image denoising, we proposed a noise removal method which combines computational harmonic analysis and variation. Firstly, we presented a nonlinear reaction-diffusion digital filter based on Nordström energy functional. For effectively overcoming speckle noise due to the reaction-diffusion process of digital filtering and the ill-posed of diffusion coefficient, we gave an improved model by introducing curvelet smoothing operator and the new diffusion function. Numerical results show that the model is not only for images with Gaussian noise, Salt&pepper noise or Speckle noise, but also suitable for mixed noise, the denoised image has higher PSNR and good visual effect.
Zhu Zhilin & Patrick Wang
Y. M. Song and X. Q. Shang, "Image Denoising Based on Reaction-Diffusion Equation and Curvelet Transform", Applied Mechanics and Materials, Vols. 40-41, pp. 554-559, 2011