ECE 285 IVR: Image and Video Restoration

This class focuses on image restoration: deblurring, denoising, super-resolution, inpainting... It covers about 100 years of techniques that have been developed up to very recent ones. In particular: spatial filters, neighborhing filters, linear vs non-linear filters, non-local means, spectral filters and Fourier analysis, heat equation, anisotropic diffusion, gradient descent, total-variation, MMSE, MAP, LMMSE, Wiener filtering, sparsity, shrinkage, wavelets, frames, lasso, proximal algorithms, patches, dictionary learning, k-SVD, EPLL, BM3D. It wont cover techniques based on deep learning, but note that the above techniques have inspired many works on deep networks and CNNs.