Abstract: Total variation regularization has been proved very successful in image restoration. However, the variational problems are hard to solve due to their nonlinearity and non-differentiability. Very recently a lot of efforts have been contributed to design fast algorithms for this kind of problems. In this talk I will present augmented Lagrangian method for total variation based image restoration models, which has been demonstrated very efficient in numerical tests.

 

About the Speaker: Dr. Chunlin Wu obtained his Ph.D from University of Science and Technology of China in 2006. After taking several postdoc positions, he joined Nankai University in 2012 as an associate professor. Dr. Wu's research interestes are in mathematcial theory and methods in image processing and computational geometry.

 

Date&Time: December 29, 2014 (Monday), 10:30 - 11:30 a.m.
Location: 606 Conference Room