Variation-Based Approach to Restoring Blurred Images Corrupted by Poisson Noise

The restoration of blurred image with Poisson noise is investigated. According to the MAP estimation of the original image, we build a new criterion to measure the fidelity of the estimated image to the original image corrupted by Poisson noise, and construct a new variational model with a regularization term. The choice of the edge-preserving regularization function is addressed. To solve the variational model, we transform it into a nonlinear diffusion equation. Numerical experiments demonstrate that the proposed method results in high performance and preserves edges and reduces the Poisson noise effectively