In this paper, we consider the reconstruction problem of photoacoustic tomography (PAT) with a flat observation surface. We develop a direct reconstruction method that employs regularization with wavelet sparsity constraints. To that end, we derive a wavelet-vaguelette decomposition (WVD) for the PAT forward operator and a corresponding explicit reconstruction formula in the case of exact data....

In this thesis, a new approach is studied for inverse modeling of ill-posed problems with spatially continuous parameters that exhibit sparseness in an incoherent basis (e.g. a Fourier basis). The solution is constrained to be sparse in the transform domain and the dimension of the search space is effectively reduced to a low frequency subspace to improve estimation efficiency. The solution sub...

Consider an n−dimensional linear system where it is known that there are at most k < n non-zero components in the initial state. The observability problem, that is the recovery of the initial state, for such a system is considered. We obtain sufficient conditions on the number of the available observations to be able to recover the initial state exactly for such a system. Both deterministic and...

Graph-based dimensionality reduction (DR) methods play an increasingly important role in many machine learning and pattern recognition applications. In this paper, we propose a novel graph-based learning scheme to conduct Graph Optimization for Dimensionality Reduction with Sparsity Constraints (GODRSC). Different from most of graph-based DR methods where graphs are generally constructed in adv...

This paper investigates new detection methods for low data-rate Machine Type Communication. Special attention is paid to uplink communication in networks where nodes are only active occasionally, which requires a joint activity and data detection at the receiver. This paper shows that a modified version of Successive Interference Cancellation is a viable approach for the detection of these low ...

EEG/MEG devices record external signals which are generated by the neuronal electric activity of the brain. The localization of the neuronal sources requires the solution of the neuroelectromagnetic inverse problem which is highly ill-posed and ill-conditioned. We provide an iterative thresholding algorithm for recovering neuroeletric current densities within the brain through combined EEG/MEG ...

We investigate the problem of Gaussian Markov random field selection under a non-analytic constraint: the estimated models must be compatible with a fast inference algorithm, namely the Gaussian belief propagation algorithm. To address this question, we introduce the ?-IPS framework, based on iterative proportional scaling, which incrementally selects candidate links in a greedy manner. Besides...