Compressed Sensing Reconstruction of an Audio signal using OMP
Shwetha A. Gangannawar and Saroja V. Siddmal
Abstract
Compressive sensing (CS) is an evolving technique for data acquisition that promises sampling a sparse signal from a far fewer measurements than its dimension. Compressive sensing enables a potentially large reduction in the sampling and computation cost for sensing signals that have sparse representation. The signal having sparse representation can be recovered from small set of linear, non-adaptive measurements. This paper mainly focuses on fixing threshold value for proper reconstruction of an audio signal. An audio signal is better reconstructed for the threshold value between (-0.02 to +0.02). Various performance parameters are measured which describe exact reconstruction of the signal. For proper reconstruction Orthogonal Matching Pursuit algorithm is used.
Keyword
Compressive sensing, sparsity, measurement matrix and OMP.