- Objective:
Blind Image Deconvolution (BID) is the problem of restoring an image X from its blurred and noisy version Y when the blur K is unknown. Mathematically, the blurred image y is modeled as a convolution of the latent image x and the blur K as
Y = K * X + ε
where ε is the noise, usually additive white Gaussian noise (AWGN). In this project, we will propose a high-efficiency blind image deconvolution model that leans on deep learning techniques for restoration processes.