作者: He Yongyang , Ding Jianyu , Hu Dong
DOI:
关键词:
摘要: The invention provides an image multi-level wavelet full sub-band compressed sensing encoding algorithm for performing sparse representation on natural images, and transforming the images into a domain with great sparsity, wherein low frequency is very important reconstructing importance of each element descendant thereof in high coefficient aredifferent from sparsity different layer coefficients, weight matrix designed according to coefficients vector, corresponding set relationship that energy times as large energy, improved measurement obtained, finally observation performed vector obtain measured value. At reconstruction end, composed HF, LH HH sub-bands, inverse transformation reconstructed image.