Talks and Poster Presentations (without Proceedings-Entry):
G. Babazadeh Eslamlou, N. Görtz:
"Binary Graph-Signal Recovery from Noisy Samples";
Poster: Signal Processing with Adaptive Sparse Structured Representations (SPARS 2017),
We study the problem of recovering a smooth graph signal from incomplete noisy measurements, using random sampling to choose from a subset of graph nodes. The signal recovery is formulated as a convex optimization problem. We reformulate the optimization problem in a way that the optimality conditions form a system of linear equations which is solvable via Laplacian solvers. We use an incomplete Cholesky factorization conjugate gradient (ICCG) method for graph signal recovery. Numerical experiments validate the performance of the recovery method over real-world blog-data of 2004 US election.
Electronic version of the publication:
Created from the Publication Database of the Vienna University of Technology.