Image fast reconstruction algorithm based on improved orthogonal matching pursuit

Sampling interval required for traditional sampling is small,when the signal is too long,the sample volume is increased,the more time is consumed,the more storage space is taken up.A improved orthogonal matching pursuit(OMP)algorithm based on atomic matching criterion of Dice coefficient(DOMP)is used to reconstruct images based on Daubechies wavelet transform base as sparse transform base and Gaussian random matrices as measurement matrices.In the atomic matching of DOMP algorithm the arithmetic mean is introduced to replace the geometric mean,the role of an important component in the vector is more prominent using of Dice measurement,so it enables more accurately to pick out the best match with the residual vector atoms from the dictionary.The experiment results show that the reconstruction error of images has great improvements under the same number of measurement values.The quality of the reconstructive image is improved about 2%,the probability of the reconstructive image is improved about 7%,this algorithm has higher application value.