Compressive Imaging Algorithm Combined the Low Dimensional Manifold Property of Image Patch with the Sparse Representation of Analytic Contourlet

Based on global sparse representation of image and local property of the patch, an efficient compressive imaging algorithm is proposed, which combined two priors: the low dimensional manifold property of local image patch and the sparse representation of analytic contourlet. The iterative hard threshold and manifold projection method are used to reconstruct images. To reduce the computational complexity, the union of a group of linear sub-manifolds is used to approximate the nonlinear manifold which tiling the whole space of patch. The initial classification is obtained based on the dominant orientation of the local image patch, then the base of every linear subspace is obtained by sparse orthogonal transform over the blocks corresponding to each class. Experimental results show that the proposed algorithm can reconstruct an image more efficiently both in the Peak Signal-to-Noise Ratio (PSNR) and visual quality than the current algorithms.