Fast Hartley transform (FHT)based reconstruction algorithm of compressed sensing

Abstract

Based on the principle of compressed sensing theory and one of the reconstruction way—Matching Pursuit (MP) algorithm, this paper improve the MP algorithm, this method analyses the time-frequency parameter property of the Gabor atom and converts the inner product calculations in MP algorithm into cross-correlation calculations to reduce the computation cost. For real signal, cross-correlation calculation can be fast done by fast Hartley transform. Test shows that this approach can greatly improve the speed of reconstruction algorithm of compressed sensing