Denoising as a Building Block: Theory and Applications in Image Processing

Explore the fundamentals and applications of image denoising in this 59-minute virtual seminar presented by Peyman Milanfar from Google Research. Delve into the theory behind denoising techniques, their importance in modern image processing, and their potential as building blocks for broader applications. Examine the properties of general nonlinear denoisers, WMap denoisers, and MAP estimates. Investigate the gradient of energy, decomposition and recomposition methods, and the interpretation of good denoisers. Learn about the use of denoisers as regularizers for general inverse problems. Engage with examples, including the Gaussian noise case, and participate in a Q&A session addressing assumptions, Jacobian symmetry, and alternative interpretations of denoising techniques.