A Committee Machine Approach for Compressed Sensing Signal Reconstruction

摘要: Although many sparse recovery algorithms have been proposed recently in compressed sensing (CS), it is well known that the performance of any algorithm depends on parameters like dimension signal, level sparsity, and measurement noise power. It has observed a satisfactory requires minimum number measurements. This different for algorithms. In applications, measurements unlikely to meet this requirement scheme improve with fewer significant interest CS. Empirically, also underlying statistical distribution nonzero elements which may not be priori practice. Interestingly, can degradation these cases does always imply complete failure. paper, we study scenario show by fusing estimates multiple algorithms, work principles, signal recovery. We present theoretical analysis derive sufficient conditions improvement schemes. demonstrate advantage methods through numerical simulations both synthetic real signals.