Plug-In Stochastic Gradient Method

BASP 2019, 2018


Plug-and-play priors (PnP) is a popular framework forregularized signal reconstruction by using advanced denoisers within aniterative algorithm. In this paper, we discuss our recent online variantof PnP [1] that uses only a subset of measurements at every iteration,which makes it scalable to very large datasets. We additionally presentnovel convergence results for both batch and online PnP algorithms.

Download paper


Y. Sun, B. Wohlberg, and U. S. Kamilov, “Plug-In Stochastic Gradient Method,” Proc. International Biomedical and Astronomical Signal Processing Frontiers Workshop (BASP 2019) (Villars-sur-Ollon, Switzerland, February 3-8), p. 75.