Recent approaches to processing and restoration of images and video have brought together several powerful data-adaptive methods from different fields of work. Examples include Moving Least Square (from computer graphics), the Bilateral Filter and Anisotropic Diffusion (from computer vision), Boosting and Spectral Methods (from Machine Learning), Non-local Means (from Signal Processing), Bregman Iterations (from Applied Math), Functional Gradient Descent, Kernel Regression and Iterative Scaling (from Statistics). These approaches are deeply connected, and their use has led to the present state of the art in imaging applications. In this talk, I will present a practical and unified framework for understanding some common underpinnings of these methods. This leads to new insights and a broad understanding of how these diverse methods interrelate. I will also discuss the statistical performance of the resulting algorithms, illustrate connections between these techniques and empirical Bayes procedures.
Peyman Milanfar received the BS degree in electrical engineering and mathematics from the University of California, Berkeley, and the MS and PhD degrees in electrical engineering from the Massachusetts Institute of Technology. Until 1999, he was at SRI International, Menlo Park, California, and a consulting Professor of CS at Stanford. He won a US National Science Foundation Career award in 2000, and the best paper award from the IEEE Signal Processing Society in 2010. He is on leave from his position as professor of electrical engineering at University of California, Santa Cruz. He is currently a visiting scientist at Google-X, where he works on Google's Project Glass, among other things. He is a fellow of the IEEE.
Cite this work
Researchers should cite this work as follows:
Birck Nanotechnology Building, Room 1001, Purdue University, West Lafayette, IN