Search

Seismic data usually show irregular spatial sampling because of cable feathering (for marine acquisitions) or physical obstacles in acquisition area (for land surveys). Seismic traces can be also irregularly distributed because of missing or noisy recordings and sensor faults. In recent literature, matrix and tensor rank optimization have been applied to achieve seismic data interpolation and to attenuate unstructured additive noise. In fact, low-rank components can capture the local features of the recorded data, such as envelope and slopes of the events. In this work, we derive a novel interpolation technique based on the low-rank approximation of matrices and tensors, which can interpolate irregularly sampled seismic data onto an arbitrary output geometry. Results on real data prove the feasibility of the proposed approach.

In this paper we present a new local-based face recognition system that combines weak classifiers to create a robust system able to recognize faces in presence of either occlusions or large expression variations. The method relies on sparse approximation using dictionaries built on local features. Experiments on the AR database show the effectiveness of our method, which achieves better performance than those obtained by the state-of-the-art l1 norm-based sparse representation classifier (SRC).

This paper provides a new regularization method for sparse representation based on a fixed-point iteration schema which combines two Lipschitzian-type mappings, a nonlinear one aimed to uniformly enhance the sparseness level of a candidate solution and a linear one which projects back into the feasible space of solutions. It is shown that this strategy locally minimizes a problem whose objective function falls into the class of the lp-norm and represents an efficient approximation of the intractable problem focusing on the l0-norm. Numerical experiments on randomly generated signals using classical stochastic models show better performances of the proposed technique with respect to a wide collection of well known algorithms for sparse representation.

Analogously to the well known greedy strategy called Orthogonal Matching Pursuit (OMP), we present a new algorithm to solve the sparse approximation problem over redundant dictionaries where the input signal is restricted to be a linear combination of k atoms or fewer from a fixed dictionary. The basic strategy of our method rests on a family of nonlinear mappings which results to be contractive in a interval close to zero. By iterating contractions and projections the method is able to extract the most significant components also for noisy signal which subsumes an ideal underlying signal having sufficiently sparse representation. For reasonable error level, the fixed point solution of such a iterative schema provides a sparse approximation containing only the nonzero terms characterizing the unique sparsest representation of the ideal noiseless sparse signal. The heuristic method so derived has been applied both to synthetic and real data. The former was generated by combining exact signals drawn by usual Bernoulli-Gaussian model and Gaussian noise; the later is taken by electrocardiogram (ECG) signals with application to the dictionary learning problem. In both cases the proposed method outperforms OMP method both regarding sparse approximation error and computation time.

The analysis of temporal connectivity graphs associated to real mobility traces in structureless Delay Tolerant Networks (DTN) has revealed the existence of a portion of nodes, called hubs, appearing frequently in optimal paths connecting arbitrary node pairs. Based on these findings, hub-based routing has become one of the most promising strategies in order to successfully deliver messages in DTNs. Considering human mobility traces, we empirically prove that the selection of a suitable set of hubs as relays, not only ensures the same probability of successful delivery, but also leads to a decrease in the mean shortest path length, combined with a limited increase of the delivery time with respect to the optimal case. By analyzing the empirical distributions of path length and travel times, we study the conditions for the best trade-off between delay-optimal and delay-sub-optimal shortest paths, which uniquely depends on the number of hubs enabled as forwarders. This analysis turns out to be useful in the design of energy-saving routing algorithms based on epidemic propagation. We found that in some real experiments, using a very little set of hubs (under 50%) and limiting the number of hops per packet to three (corresponding to the mean delay-optimal shortest path length), epidemic routing guaranteed a delivery ratio around 90%, when tolerating at most double travel time only for a little fraction of messages.