LDCT image restoration method based on unpaired data

The invention discloses an LDCT image restoration method based on unpaired data, and the method comprises the following steps: S1, inputting a pre-training set comprising a plurality of unpaired LDCTimages and a plurality of unpaired HDCT images into an image restoration network so as to pre-train the image restoration network, wherein the image restoration network comprises two generators and two discriminators, and each of the two generators comprises a convolution layer and a residual channel attention block; S2, adjusting the generators in the image restoration network by using an adjustment training set comprising a plurality of paired LDCT images and a plurality of paired HDCT images to obtain a trained generator; and S3, inputting an LDCT image to be recovered into the trained generator to obtain a high-quality LDCT image, wherein the number of the LDCT images and the number of the HDCT images in the adjustment training set are lower than the number of the LDCT images and the number of the HDCT images