Low-dose CT reconstruction method combining prior image and convolutional sparse network

The invention discloses a low-dose CT (Computed Tomography) reconstruction method combining a prior image and a convolutional sparse network, and belongs to the technical field of computed tomography. The method comprises the following steps of: firstly, acquiring a high-quality prior image by adopting a distinctive feature representation method; secondly, designing a feature fusion convolutional sparse network; next, calculating an error between the reconstructed image and the prior image, and estimating an iterative reconstruction gradient according to the error; then, carrying out fine tuning on the reconstructed image by adopting total variation constraint; and finally, iterative reconstruction of the low-dose CT is realized in a module cascade form. The algorithm comprehensively utilizes the advantages of the prior image planned to be scanned and the convolutional sparse network, has the advantages of being high in interpretability, high in reconstructed image quality, few in noise artifacts and the like