In this work we propose a new image smoothing and edge detection technique that employs a combination of nonlinear diffusion and bilateral filtering. The model is based upon two very well established methodologies in the image processing community, which makes the method easy to understand and implement. Our numerical experiments show that the proposed model is capable of achieving more accurate reconstructions from noisy images, as compared to two other popular nonlinear diffusion models in the literature. We also propose a new and simple diffusion stopping criterion, based on the second derivative of the correlation between the noisy image and the filtered image. This indirect measure allows stopping the diffusion process very close to the point of

maximum correlation between the noise-free image and the reconstructed image, in the absence of the former. The stopping criterion is sufficiently general to be applied with most nonlinear diffusion methods normally used for image noise removal.